This is the old United Nations University website. Visit the new site at http://unu.edu
Abstract
Introduction
Processes that promote or hinder developmental competence
Implications for intervention
What not to do
Principles for effective and cost-effective interventions
Acknowledgements
References
Theodore D. Wachs
The author is affiliated with the Department of Psychological Sciences at Purdue University in West Lafayette, Indiana, USA.
Although the reduction of child morbidity and the promotion of physical growth are important and necessary aspects of child development, these criteria by themselves do not define the adequacy of children’s development. There are also behavioural-developmental criteria that emphasize the promotion of competence. The competent individual is one who can effectively adapt to and interact with his or her environment. Traits that define individual competence fall into five domains: cognitive skills, temperament/ personality, motivation, self-perceptions, and interpersonal style. These domains are not completely independent, and there is at least partial overlap. The expression of individual differences in competence is partially moderated by context. Further, not all children achieve competence. Over time some children fall further and further behind their peers in their developmental course. In under standing what biological and psychosocial factors influence the development of individual differences in competence, four principles are critical.
First, most aspects of individual competence are multidetermined. This means that interventions designed to facilitate development must be multifocal in nature, integrating influences from different domains. Second, influences upon children’s development tend to be specific in nature. This emphasizes the importance of targeting specific interventions to specific outcomes. Third, individual developmental influences rarely operate in isolation from each other. Developmental risk factors tend to cluster together, as do developmentally protective influences. The extent of the impact of a given developmental risk factor will depend, in part, on the degree to which this risk factor covaries with other risk factors. Fourth, developmental risk and protective factors operate across time. Early exposure to developmental risks may increase the individual’s susceptibility to later risk factors (sensitizing) or may limit the degree to which the individual can profit from later exposure to protective factors such as intervention (blunting). Early exposure to developmentally protective factors may attenuate the impact of later exposure to developmental risk factors (steeling).
Principles underlying the nature and nurture of individual competence emphasize the need to use an IT-AT intervention strategy (Integrate Target Across Time). This means the need to integrate multi-domain interventions, target our intervention strategies to different contexts, risk conditions, and outcomes, and provide for recurring interventions across time to maximize the chances of long-term gains in individual competence.
When we speak of facilitating child development, particularly in less developed countries, our focus, typically, is on issues of health and physical growth [1]. Reducing child morbidity and promoting physical growth are important and necessary aspects of child development, but these criteria by themselves do not define the adequacy of children’s development. In addition to physical criteria there are also behavioural-developmental criteria that emphasize the promotion of competence. Although a standard definition of competence remains elusive, there appears to be a general consensus among developmental researchers that the competent individual is one who can effectively adapt to and interact with his or her environment. Criteria for defining effective adaptation include the ability to meet major developmental goals viewed as appropriate for a given individual at a given age in a given context, as well as coping with environmental challenges and stresses [2, 3].
The specific adaptations and interactions that define individual competence are context-specific. What defines a competent pre-school child in rural Kenya (collecting firewood, caring for younger siblings) will be much different from what defines competence in a pre-school child in the United States (development of symbolic play skills) [4]. Similarly, the criteria that de- fine competence in a traditional culture may be less appropriate in a culture undergoing major transitions, such as industrialization or urbanization [1, 5]. For example, an important yet little-known study by Albizu-Miranda et al. (cited in Heber and Dever [6]) showed that the ability of individuals with below-average intelligence to succeed economically dropped radically as the level of urbanization-industrialization of communities increased.
Although the question of whether or not an individual is regarded as competent is context-specific, there also appear to be dimensions of individual competence that cut across different contextual settings. Evidence for this assumption is seen in a study in which clinicians in eight less developed countries in Asia, Africa, and the Caribbean were asked to define the criteria they would use to judge a child as being mentally retarded [7]. The same five behavioural dimensions were used by clinicians across all cultures, although there were cross-cultural differences in the degree to which they rated the importance of each of them. If there are common behavioural-developmental dimensions underlying individual competence, what are these dimensions? Typically we have defined competence primarily in terms of intelligence. Although intelligence is a critical component of competence, other behavioural domains may be equally important [8]. To understand what are the critical domains underlying competence in children, I will turn to studies of “resilience” [9] or “positive deviance” [10]. Both of these terms refer to children born into situations involving multiple cumulative biological (e.g., malnutrition, chronic morbidity) and psycho-social (e.g., poverty, lack of environmental stimulation, physical abuse) risks, who nonetheless manage to function at a level well above what we would expect given their background. As described by Werner and Smith, these are children who, in spite of the worst that life could throw at them, developed into adults who “worked well, played well, loved well, and expected well” [11]. Many of the factors that make children resilient involve aspects of the biological and psychosocial context (e.g., greater intake of foods of animal origin, smaller family size with longer birth spacing, secure attachment to caregivers, adults who are available to the child in times of crisis or stress [10, 11]). However, it is also clear that resilient children are likely to possess a set of individual characteristics that allow them to receive a major share of available physical and psychosocial resources, even in environments where such resources are, for the most part, lacking [12]. I would argue that the characteristic traits that produce resilience in children growing up under such extreme conditions are also likely to be the characteristic traits that underlie the development of competence in children living in less extreme situations.
What are the individual traits that underlie competence? As can be seen from table 1, the various traits that define resilience can be said to fall in five domains: cognitive skills, temperament and personality, motivation, self-perception, and interpersonal style. Although they are not listed in table 1,1 would also note the importance of physical characteristics. Infants and children judged to be unattractive by their caregivers or peers are more likely to be rejected and receive less support than children judged to be attractive [13-15]. To some extent, competence can also be influenced by other physical characteristics, such as race or sex, although for these traits competence will usually be a function of cultural or contextual bias against individuals of a particular race or sex.
TABLE 1. Individual traits associated with resilience
|
Domain |
Example |
|
Cognitive skills |
Alertness (infancy) |
|
Temperament/personality |
High activity level (infancy) |
|
Motivation |
Need for competence |
|
Self-perceptions |
Secure attachment (infancy) |
|
Interpersonal style |
Ability to use adult as resource |
There is research showing how intelligence develops and differentiates across the life span [18] and how early temperament maps onto later personality dimensions [19]. However, such descriptive information would be relatively limited. Rather than focusing on descriptive changes in the various traits and domains underlying children’s competence, I will focus on three underlying principles about the nature and development of competence and then on the more fundamental question of what promotes competence in children.
Competence is partially moderated by context. Specific skills that serve to promote competence in one context may be irrelevant or even detrimental to the development and expression of competence in other contexts. For example, high activity levels, which would be appropriate for children living in nomadic tribes, become highly maladaptive when such children are placed in a school setting [20]. Similarly, the overlapping communication patterns that define the competent Hawaiian child in his or her family are viewed as developmentally inappropriate when such children are faced with the demands of traditional Western schooling [21]. On the other hand, it is important to keep in mind the caveat of Diaz-Guerrero [22] that in traditional societies undergoing modernization, those individuals who achieve most may be those who deviate to some degree from traditionally valued coping skills. Thus, our focus must be on those behaviours within a given domain that allow an individual to adapt to his or her current context, but do not hinder adaptation if the current context changes.
Not all development is progressive. Although we typically think of children’s behavioural development as progressive in nature, moving towards higher levels of competence as the child gets older, such progression is not necessarily a given. The psychological analogues to physical growth concepts such as stunting are seen in terms such as cumulative deficit [23, 24] and negative-risk feedback cycles [25, 26]. These terms refer to the fact that some children fall further and further behind their peers in their developmental course, with the likelihood of catching up declining the longer development proceeds. One example of regressive development is seen for children with early-appearing antisocial behaviour patterns who, over time, become locked into progressively wider negative interactions, first with their parents, then with their peers, and finally with school authorities. The end result is a chronic pattern of antisocial behaviour that continues into and through adulthood [27]. Unless there are major contextual changes, such children are essentially locked into negative developmental cycles. However, the likelihood of major contextual changes is often substantially reduced, given that the behaviour of these children acts to close off alternative contextual niches that could serve to redirect their developmental trajectory [28].
Traits defining competence covary. I have defined competence on the basis of individual traits in different domains. This reflects a bias of developmental researchers, who tend to focus on specific areas of development taken in isolation. In reality, the domains that define competence in children rarely fall into such nice neat categories. Rather, many aspects of children’s development from supposedly different domains share common elements [1]. Some shared elements may be definitional. For example, definitions of intelligence used in many non-Western countries encompass not only traits typically thought of as intellectual in nature (such as memory or verbal facility), but also traits that are considered interpersonal in nature (such as politeness and respect for elders) [29]. For other traits, covariance is inherent in the nature of the traits themselves. The behavioural characteristics defining self-regulation involve linked contributions from the domains of temperament and cognition [30]. Similarly, the ability to understand the perspective of other individuals involves both social and cognitive components [31], whereas the ability to react empathetically to others involves affectual and cognitive components [32]. Rather than the different domains defining competence having sharply defined boundaries, the concept of “fuzzy” borders maybe more applicable [33] when attempting to distinguish between traits from different domains.
Although there is covariance across the different traits and domains defining competence, such covariance does not mean that differently labeled domains or traits refer to the same thing. Evidence for this is seen in a study of children from high-risk backgrounds who were characterized by teacher reports and school records as being resilient during periods of stress. Although these children were all resilient in some domain of competence, such as intellectual or social skills, only 15% of these children were found to be resilient in more than one domain [34].
The concept of covarying but not isomorphic domains of competence is illustrated in figure 1 using three domains: temperament, cognition, and motivation. Although there is a small degree of overlap among all of these domains (central shaded area) and a larger degree of overlap between any two domains (off-central shaded area), each domain also has its own unique properties that are not shared by the others. The nature of relations among the different domains underlying competence has certain implications for interventions designed to promote competence. For maximum impact we would want to target our interventions at the point where there is a clear intersection of each of the relevant domains. However, such targeting requires a level of measurement accuracy and conceptual knowledge that we simply do not possess at present. Hitting the overlap point is more likely a matter of luck than of design. In practice this means that the impact of specific unidimensional interventions that promote competence in a single domain may not necessarily generalize to other domains of competence.
FIG. 1. Covariance among different domains of development
Development is multidetermined
Specificity
Covariance among developmental influences
Temporal moderation
Rather than discussing the specific contributions of different types of developmental influences, such as nutrition, morbidity, psychosocial environment, or culture, this section will focus on the processes underlying the contributions of these influences to the development of competence, with discussion of four principles: multiple determinants, specificity, covariance, and temporal moderation.
In understanding the effect of specific influences upon individual behavioural-developmental variability, behavioural geneticists emphasize genetic contributions, nutritionists emphasize nutritional contributions, psychologists emphasize environmental contributions, and anthropologists emphasize cultural contributions. The traditional goal within each discipline is to isolate the unique contributions to development of influences from one’s chosen field of study. Potentially relevant influences from other fields all too often tend to be regarded as nuisance variables that hinder our ability to isolate the unique contributions of what we are focusing on. However, parsimony notwithstanding, the factors that influence individual developmental variability rarely operate in isolation from each other, and most aspects of individual behavioural development are multidetermined. Many specific individual influences are necessary, but few are sufficient as an explanation.
Let me support this conclusion in several ways. First, let us deal with those cases where intervention based on a single targeted influence can have a major impact upon development. The case of prenatal administration of folate or iodine as a means of preventing neural tube defects [35] or cretinism [36] documents the power of specific single targeted interventions. However, such cases tend to be the exception rather than the rule [37]. For example, although there are many single-gene disorders that can have a major impact upon development, these disorders tend to be relatively rare [38]. Even in a situation where we are dealing with a single influence that has a major impact upon behaviour and development (such as a rare genetic defect), this does not rule out the possibility of important contributions to the development of affected individuals from influences in other domains. For example, research has illustrated the contributions of environmental influences to the development of individuals born with rare sex chromosome anomalies [39] or with rare genetic disorders such as the cri du chat syndrome [40].
The contribution of multiple influences to individual behavioural developmental variability can be illustrated by the nature of genetic contributions to development. As shown in figure 2, the path from genes to outcomes is indirect, circuitous, and complex. Genes themselves directly affect only microbiological processes - essentially the transmission of DNA to RNA [41]. Thereafter, we have a complex multilevel pathway, including contributions from both biological variables (e.g., nutrition, teratogens) and environmental stressors and supports. Not only is this pathway complex, but it is also bi-directional. Thus, whether structural genes are actually expressed or not depends on the action of regulatory genes, which are directly sensitive to a variety of non-genetic influences such as hormonal levels or nutritional status [42].
FIG. 2. Pathway from genes to development
Another example of the role of multiple influences that is of particular relevance to children at risk in both developed and less developed countries is school failure and drop-out rates. [1, 43, 44]. Level of school attendance has been associated with higher cognitive performance in both developed [45] and less developed countries [46]. Level of schooling facilitates cognitive gains associated with nutritional supplementation [47]. Particularly in less developed countries, the level of maternal education is negatively related to frequency of births [48] and is positively related to children’s level of physical growth [49], children’s survival rate [50], and quality of maternal rearing practices [51].
A summary review of some of the many factors contributing to school failure school drop-out is shown in table 2. It is clear that school failure or drop-out can result from any number of multiple influences combining in any number of ways. All too often, these influences cumulate over time, producing a progressive disengagement from school [43].
Obviously not all of the variables shown in table 2 will be of equal importance in influencing the probability of school failure or school drop-out. However, the lesson that can be drawn from the data shown in table 2 is that when multiple influences are operating, the chances of having a major impact upon school failure or school drop-out, as a function of changing just a single influence, are not particularly high. Indeed, going beyond just school failure, what the evidence consistently indicates is that significantly better prediction of individual behavioural- developmental variability occurs when developmental influences from multiple domains are looked at in combination rather than in isolation [47,59-64]. The reverse conclusion also holds. If potentially salient influences from domains other than the one we are targeting are not taken account of, the results may well be outcomes that are either non-significant or opposite to what is desired. This is demonstrated most dramatically in a study by Grantham-McGregor and colleagues [65], showing that the impact upon school performance of feeding breakfast to undernourished children varied as a function of school context. As seen in figure 3, in an organized, non-crowded school environment (school A), breakfast feeding was shown to facilitate school performance. In non-organized, chaotic school environments, breakfast feeding either had no effect (schools B and D) or had a potentially detrimental impact (school C), perhaps energizing children to become more reactive to chaos in the classroom.
There is increasing evidence that many developmental influences tend to act in a relatively specific fashion, impacting only upon a restricted set of developmental outcomes [66]. In many cases different aspects of development are predicted by entirely different developmental influences. For example, delayed toddler mental development is equally well predicted by a combination of biomedical and environmental risks; delayed toddler motor development is best predicted by biomedical risks; delayed toddler receptive language skills are most strongly predicted by environmental risks [67]. Other examples of this phenomenon, which I have called “specificity,” are shown in table 3.
TABLE 2. Factors influencing school failure and school drop-out
|
Individual characteristics |
Contextual characteristics |
School characteristics |
Interactive influences |
|
Gender |
Family socio-economic status |
Educational quality of school environment (e.g., well
organized, teacher’s time spent teaching, feedback provided to
students) |
Culture by gender |
|
Ethnicity Intellectual level |
Family educational level |
School size |
Culture by ethnicity |
|
School-related behavioural disorders (e.g., attention deficit
disorder-hyperactivity, learning disability) History of school failure |
Quality of home environment |
Educational expectations of school |
Fit between characteristics of culture and school
environment |
|
Nutritional status |
Parental expectations |
|
|
|
Morbidity status |
Educational support from peer culture |
|
|
|
Destructive behaviour patterns |
|
|
|
|
Adolescent sexual habits |
|
|
|
|
Identification with school environment |
|
|
|
The fact that the impact of individual developmental influences tends to be relatively specific means that we cannot assume that a given influence that changes developmental patterns for certain outcomes will necessarily work for a different set of developmental outcomes. Generalizability of the effects of specific developmental influences is something that must be tested rather than something we can assume on faith. A specific developmental influence that has facilitative impact upon behaviour and development in one outcome domain maybe irrelevant to outcomes in a second domain and may, under some circumstances, hinder development in a third outcome domain. For example, although it has been assumed that high levels of parental responsivity to children’s verbalization and interaction patterns will impact in a positive way upon most aspects of development, available evidence suggests that some aspects of parental responsivity (caregivers responding non-verbally to children’s vocalizations) act to inhibit rather than facilitate development [66].
FIG. 3. Effects of nutritional supplementation on school behaviour in different school contexts. School A was organized and not crowded. Schools B-D were disorganized and chaotic.
The operation of specificity emphasizes the importance of focusing on unique patterns of relations between specific developmental influences and specific outcomes. However, before narrowing our focus too much, we need to consider the operation of covariance among multiple developmental influences. Covariance refers to the fact that in nature, individual developmental influences rarely operate in isolation. Rather, combinations of two or more developmental influences often co-occur at a greater than chance probability. Specific developmental risk factors tend to cluster together, as do specific developmentally protective influences [75]. For example, if an individual is deficient in vitamin B6, there is a greater than chance likelihood that the individual is also deficient in other B vitamins [76]. Similarly, there is ample evidence documenting the covariance between inadequate psychosocial rearing environments and deficits in general nutritional status [77, 78].
The covariation among developmental influences from different domains may be inherent in the nature of the influences themselves, as seen in the bidirectional covariance between individual malnutrition and increased risk of morbidity, where malnutrition depresses immune system functioning while morbidity reduces appetite [79]. Alternatively, covariance may be the result of differential treatment of individuals with different characteristics, as occurs for adolescents with antisocial behaviour patterns who have a higher probability of eliciting rejection and anger from their parents [80]. Although covariance among developmental influences is a well-established fact, what must be kept in mind is that covariance is always probabilistic in nature [37]. For example, although there is an elevated risk of chronic marital discord in families where a parent is mentally ill, in the majority of such families we do not find elevated levels of discord [81]. Similarly, although infants with a “difficult” temperament are at elevated risk for evoking parental hostility and rejection, many difficult infants have excellent relations with their parents [66].
TABLE 3. Examples of specificity of developmental influences
|
Influence |
Evidence |
Refs. |
|
Central nervous system |
Performance deficit varies as a function of what central nervous system
area is damaged |
68,69 |
|
Biomedical influences |
Differential impact upon physical versus behavioural development depends
upon which type of teratogen the individual is exposed to |
70 |
|
Nutritionally related growth markers |
Different anthropometric patterns at birth related to different types
of adult medical disorders later in life |
71 |
|
Proximal environment |
Maternal vocalization patterns and level of maternal response to distress
differentially predict toddler language and emotionality |
72 |
|
Influence of parental rearing style upon change in adolescent competence
varies as a function of what cognitive areas and what types of parental
rearing styles are assessed |
73 |
|
|
Demographic-cultural influences |
Strong family religious beliefs inhibit adolescent behavioural problems
but do not influence adolescence academic competence, which is uniquely
predicted by family income |
74 |
First, the stronger the covariance among multiple developmental influences, the more likely we are to find variability in outcomes. In this case the generalizability is not due to a general impact of a single influence but rather to multiple specific impacts associated with multiple covarying influences. For example, the generalized effects associated with poverty or community disorganization may reflect the fact that these terms refer to aggregated combinations of multiple and specific risk factors [82].
Second, when covariance exists among multiple developmental influences, a traditional approach has been either to ignore existing covariances or to control for such covariance statistically or experimentally as a means of isolating the unique contribution of a specific predictor [75]. Although it is relatively easy to isolate a single influence statistically or experimentally, such isolation may make little sense in a real-world setting, where individuals are simultaneously exposed to multiple covarying risks and protective factors. The operation of covariance leads to the suggestion that the appropriate unit of analysis should not be an individual developmental influence taken in isolation, but rather the covarying pattern of linkages among multiple influences [64]. One way of doing this is through a technique known as pattern analysis, through which individuals with similarly covarying biological individual and psychosocial characteristics are grouped into a specific cluster. Rather than individual influences, cluster membership is used to predict developmental outcomes [83]. Stronger predictions may occur when clusters of covarying influences are used as the unit of analysis rather than individual influences taken in isolation. For example, the impact of a single protective environmental factor upon individual development will depend upon the degree to which this factor covaries with other protective environmental factors encountered by the individual [84].
To understand the sources of variability in individual behavioural development, we must deal not only with multiple influences, many of which both covary and are highly specific in their impact, but also the fact that both development and the role of developmental influences operate across a background of time. Some aspects of temporal moderation are relatively well known. The concept of sensitive periods is one such aspect. The concept of sensitive periods refers to the idea that the impact of specific developmental influences will vary as a function of the age of the individual. Across multiple disciplines, we find the common assumption that developmental trajectories are most sensitive to modification by extrinsic influences during the period when the individual is most rapidly developing [85,86]. For example, there is evidence that the central nervous system is more sensitive to the impact of injury or environmental toxins and more able to compensate after exposure to injury or toxins early in life, when central nervous system structures are still maturing [87]. The same assumption underlies the emphasis placed on supplementing growth-retarded infants during the first year of life while physical growth is at its maximal rate [88]. This assumption has led many researchers to emphasize the importance of early biological and psychosocial intervention, since the early years are presumed to be a period of maximal growth and development [89].
Early developmental influences can act to moderate the impact of later developmental influences in three ways. First there is sensitization, which refers to prior developmental influences making the individual more sensitive to later stressors [90]. For example, children with a history of malnutrition are more likely to be sensitive to the detrimental impact of subsequent short-term nutritional stress than children without such a history [91,92]. Similarly, children who have a history of prior developmental problems are more likely to be at long-term risk when faced with later stressors, such as parental divorce, than those without such a history [93]. Both infrahuman [94] and human data suggest that long-term sensitization may be due to relatively permanent changes in individual physiological characteristics associated with exposure to early developmental risks [71,95].
Second, although less evidence is available, early developmental influences may also act as a moderator through a process called steeling, wherein early influences act to protect the individual against the detrimental impact of later stressors [96]. There is a long history of infrahuman studies showing how pre-weaning exposure to mild environmental stresses acts to make organisms more stress resistant in adulthood [97]. At the human level, there is evidence that children with a secure attachment are more likely to show competent responding when faced with later environmental challenges than children who are insecurely attached [98].
Finally, there is also moderation as a function of blunting, namely, prior exposure to risk influences making the individual less able to benefit from subsequent facilitative developmental influences. For example, children with poorer early nutritional status were less able to benefit from later rearing in highly advantaged circumstances than more adequately nourished children from similar backgrounds [99-101]. Similarly, children who were either institutionally reared or reared in highly disorganized family environments were less able to benefit from the impact of either later intervention or later adoption into a more advantaged family environment [102-104].
Although there clearly is something unique and important about the role of developmental influences early in the life span [66, 105, 106], the complexity of temporal processes does not lend itself to a recommendation to focus intervention efforts just in the early period of life. Contradicting this recommendation are infrahuman studies suggesting not only that the central nervous system may be sensitive to the impact of environmental influences through late adulthood (well beyond the period of maximal central nervous system development) [107], but also that certain biological stresses, such as malnutrition occurring early in life, may act to extend the period of maximal central nervous system development, thus allowing for greater catch-up time [94]. Also, human research contradicts this idea, indicating that individuals do not become less sensitive to developmental influences over time, but rather that they become sensitive to different types of developmental influences [66]. For example, physical growth in the first six months of life is uniquely sensitive to the level of nutritional intake. However, for older children the level of growth hormones and sex steroid production has a more salient influence upon later physical growth [108, 109]. The importance of going beyond just the early period of life is also seen in the fact that the impact of early biological and psychosocial risks upon later development can be partially or even totally overcome by later exposure to more facilitative biological and psychosocial influences [66]. For example, more adequate food intake and better environmental stimulation later in life can at least partially compensate for the effect of early childhood malnutrition [101,110].
Although early influences can act to moderate the impact of later influences, and later influences can act to moderate the impact of prior influences, ultimately it is important to recognize that much developmental variability is the result of a cumulative chain of developmental influences. For many aspects of development, it may well be the accumulation of influences over time that is most critical, rather than influences operating at a given point in time. The impact of cumulative influence processes can be seen in terms of the decline in active coping strategies used by children in response to chronic and continuing societal violence [111]; in the exhaustion of family resources as stresses on the family continue to occur [112]; in the greater retardation of physical growth rate found as malnutrition cumulates [113, 114]; and in the sharp increase in the risk of future gastrointestinal disorders for individuals with a history of previous gastrointestinal disorders [115]. Looking across domains, the cumulative impact of both malnutrition and lower exploration opportunities may ultimately result in children developing a passive, helpless pattern of learning in relating to their environment [47,116]. Similarly, a lack of educational support at home plus low parental expectations increases the probability of children starting off poorly prepared for school; this poor preparation in turn increases the probability of children doing poorly in the first few grades in school, which in turn increases the probability that the child will become less and less involved with school. This cumulative process of disengagement ultimately maximizes the risk of later school failure and drop-out [43].
Again, however, I must stress the probabilistic nature of developmental influences. Even where we have multiple influences cumulating towards a specific outcome, the probability of such an outcome is not necessarily 100 percent. An example is the research on women who were reared in institutions in childhood because of inadequate family environments [117]. Although as a group these women were at greater risk for later adult behaviour problems, those who had positive school experiences in their childhood had better adult adjustment than those who did not have such positive school experiences.
Although in no way denying the unique importance of the early years or of time periods of maximal growth as a focus for intervention efforts, the complex nature of temporal processes means that intervention efforts, whether biological or psychosocial, should not be restricted just to the early years or to periods of maximal developmental growth. Developmental interventions, particularly at the human behaviour level, rarely function as a form of inoculation. We cannot necessarily assume that the impact of a specific developmental influence at a given point in time will be maintained across time. A critical question is not so much whether recurring interventions are needed, but rather the question of when and how such recurring interventions should be provided.
After hearing about the complexities involved in human development, a typical response by individuals involved in public policy decisions is “If things are this complex, can cost-effective interventions really be developed?” I will respond to this question first in terms of economics and then at the level of programme development. Economically, it is important to stress that the complex picture I have painted, in regard both to development and to the nature of influences on development, is not an artifact. These complexities are real and there are potential economic consequences if we ignore them when designing intervention programmes to influence human behavioural development. Designing intervention programmes on the basis of a false assumption that human development is both simple and easily changed is a strategy that may save us initial costs, but over time for most developmental outcomes the inadequacies of such an intervention strategy will become all too apparent. There are exceptions to this (e.g., iodine supplementation), but these are truly exceptions and not the rule. As a general rule, the more a given domain of developmental competence is multidetermined, the less likely we are to find maximal gains associated with a single time-limited intervention. Multilevel repeated interventions that take account of existing developmental complexities may have a greater start-up cost, but in terms of long-term effectiveness and generalizability of effects, evidence indicates that this type of intervention will be far more cost-effective over the long run [118]. In terms of programme development, existing evidence can guide us, in terms of both what not to do and what to do.
Avoid the assumption that one type of intervention taken in isolation will apply equally well to all outcomes. For example, school curricula that foster academic achievement may do little to foster either behavioural adjustment or emotional maturity [53].
Do not expect what works in one context to generalize equally well to another context. The generalizability of successful intervention programmes can be limited by a variety of cultural factors, such as the degree of gender segregation in childhood [21], the preferred mode for teaching young children [119], and whether the culture is family or child centred [120].
Do not design interventions that focus primarily on a single outcome domain. As in the case of cognition and motivation, gains in one domain can be strengthened and stabilized if interventions also target relevant overlapping domains [1, 121]. In addition, given the operations of specificity, there is always the possibility that intervention-related gains in one domain may be offset by intervention-related losses in a different domain. For example, providing additional play objects to black South African pre-school children, while increasing their cognitive competence and their appropriate use of objects, was also found to decrease their use of language and increase their level of solitary as opposed to social play [122].
Never design interventions without first asking what existing conditions can interfere with the potential gains individuals can realize from these interventions. To the extent that interfering conditions exist, dealing with these conditions as part of the intervention may be as important as the actual intervention itself. I have already noted the study by Grantham-McGregor et al. [65] showing how chaotic school environments can compromise the impact of nutritional interventions. Other examples similarly document the need to take account of potential interfering conditions. Thus, in a population where there is a high intake of unleavened whole-grain products and tea, there may be limitations on the degree of functional benefits to be found by providing iron supplementation [123]. Focusing on providing macro- or micronutritional supplements to a family in a culture where the parents believe that disorders such as marasmus are the result of supernatural actions or of individuals not meeting religious obligations [124] or where parents believe that undernourished infants do not require any special help other than making food available [49] is a strategy that ignores existing realities that could compromise the impact of intervention. Providing home-based early environmental stimulation to at-risk infants living in crowded chaotic home contexts is a strategy that will result in limited gains, even if the interventions themselves are state of the art [125].
In evaluating the effects of intervention, do not focus just on main effect group differences. Even when there are significant group differences between individuals receiving intervention and those who are not, it is equally essential to pay attention to the level of intra-group variability within the intervention group. In all too many cases, we see a situation where a few children in the intervention group show major gains, a majority of children in the intervention group show very modest changes, and some children receiving intervention show either no change or even a regression over time [126]. In this type of situation, can we really say that we have carried out a generally successful intervention?
Integrate
Target
Across time
Many of the intervention principles that Come from an understanding of the complexity of human behavioural developmental variability can be summarized by the simple acronym IT-AT, which stands for “integrate target across time.” Integrate refers to the fact that aspects of human competence are affected by covarying influences from multiple domains. This means that we need to integrate multidomain interventions when attempting to influence the course of development. Target refers to the fact that intervention strategies need to be tailored for different cultural contexts, for different risk conditions, and for different outcomes rather than assuming that a given intervention will equally influence all outcomes for all individuals under all circumstances. Across time refers to the need for interventions to reoccur over time to maximize the chances of long-term gains.
To the extent that an outcome is determined by multiple influences, intervention strategies should also encompass multiple influences. The critical question is which influences from which domains to include in the intervention. In an ideal sense, we would first determine what domains were relevant and within each domain what specific factors were most salient for the outcomes targeted. We would then do a survey of our population to determine which specific factors from each critical domain were lacking or were in excess, and which of those that were lacking or were in excess might most easily be manipulated to promote the desired gains. Such a strategy, although theoretically correct, is far too complex to be of use in most intervention situations in less developed countries.
An alternative easier strategy that allows us to integrate across multiple influences involves building upon existing covariances among developmental influences. We are far more likely to get maximal and lasting gains if we build on existing covariances than if we ignore the covariance among developmental influences. Let us take the covariance between nutritional deficit and inadequate psychosocial stimulation as an example. Available evidence documents that we can improve the health status of young children living in populations at risk for zinc deficiency by zinc supplementation [127, 128]. Functional consequences of better child health include more regular school attendance and better attention to the environment [129]. As shown in figure 4, zinc supplementation promotes better health, and better health status in turn makes young children more responsive to the environment. To the extent that micro-nutrient deficits covary with inadequate psychosocial rearing conditions [63,95], it would be both logical and important to build on existing covariation by combining zinc supplementation with a programme of psychosocial stimulation designed to promote cognitive performance, since supplemented children will be more likely to be receptive to such stimulation than unsupplemented children. As has been shown in Asia [130], Africa [119], and the Caribbean [131], relatively low-cost, culturally appropriate psychosocial stimulation programmes can be provided in the context of other interventions designed to reduce the impact of morbidity or malnutrition [118]. This may be particularly true if we utilize available technology, such as television, which is often found even in the poorest villages, as a mechanism for delivering appropriate psychosocial stimulation to large numbers of children [132]. Integrated interventions based on covarying developmental influences will allow us to influence multiple aspects of development in a manner that is both cost effective and likely to have long-term developmental benefits.
FIG. 4. Zinc, health, and psychosocial stimulation
Targeting interventions involves issues such as what outcome we are targeting and whom we are targeting. In regard to the former question, all too often our intervention efforts are directed towards facilitating performance on a specific cognitive or behavioural measure (e.g., the Bayley) without asking the question as to whether gains on this particular outcome measure are really what is most desirable for a child in a given population. Rather than targeting for performance on available and commonly used behavioural or developmental measures, we should target three domains of competence, even when the measurement of gains in these domains is more difficult.
First, it is important to target those individual and behavioural characteristics that allow the individual to function adequately in his or her culture. Existing knowledge about cultural values will tell us, for example, whether we should target verbal or social aspects of intelligence or independence versus the ability to cooperate with others.
Second, particularly in cultures undergoing urbanization or industrialization, we should target those cognitive and behavioural characteristics that allow the individual to adjust to changing conditions. As noted previously, these are not necessarily the same characteristics that allow the individual to adapt to a traditional context. Particularly in regard to adaptability, we should focus on those contextual conditions and individual characteristics such as family size, maternal educational level, and enhanced child cognitive performance in the pre-school years that promote individual literacy [133]. Particularly in cultures where education is viewed as a valid goal for males but not for females, this may mean targeting not only the child but also the family.
Third, a final targeted dimension should be those cognitive, individual, and behavioural characteristics that allow individuals to adapt when encountering later stressors. Of particular relevance here is the promotion of less easily measured but critical individual characteristics, such as the capacity for self-regulation, the development of secure attachment, adequate self-esteem, and the ability to seek out appropriate environmental resources in times of stress. Although easily obtainable measures of many of these characteristics may not exist at present, by targeting these types of characteristics we are more likely to promote long-term generalizable gains than by targeting less critical characteristics that can be more easily measured.
In regard to the question of whom we target, in a situation where resources are finite, it seems almost too obvious to argue that it is most cost effective to focus primarily on those children who are most at risk in a given population. Unfortunately, too many of our most commonly used risk indices, such as low socio-economic status or living in an area characterized by inadequate nutrition or poverty, although valid at a group level, are relatively insensitive at an individual level. We know, for example, from the classic work of Craviotto and DeLicardie [134], that even within a low socio-economic poverty group, there are individual differences in the risk of children becoming malnourished, with quality of rearing environment rather than poverty itself being the critical distinguishing feature. Similarly, we know from the anthropological work of Scheper-Hughes [16] that infants whose individual characteristics do not fit the characteristics that their mothers feel are essential for survival are at particular risk for both malnutrition and mortality. Within a given population, those children who are most exposed to multiple biological and psychosocial stressors will be most at risk [135]. This emphasizes the importance of developing risk indexes to identify individual children who fit the criteria of high risk within a given population. Examples of relatively easily measured and valid individual characteristics that could go into such a risk index are shown in table 4. Developing a risk index to target the most vulnerable children in a high-risk population, where it is not economical to treat everyone, not only allows for more cost-effective intervention but also avoids potential problems with targeting intervention to non-risk children, as seen in the case where there was a retardation in weight gain when iron supplementation was provided to children with adequate iron status [136].
TABLE 4. Indices used to define individual risk in a high-risk population
|
Low birthweight Inadequate growth velocity in first six months Level of morbidity above population mean in first two years of life Iron-deficiency anaemia Family in which sibling spacing is closer than population average Family in which older siblings are developmentally delayed or atypical Mother uneducated Individual characteristics provide a poor fit to preferred cultural values History of early school failure |
There is no doubt that targeting early in development is very important, as is targeting during periods when specific developmental competencies are undergoing rapid change. However, just targeting early or during periods of rapid change may not be sufficient. In populations where children are exposed to multiple risks, even targeting during a sensitive time period may be ineffectual if the interventions are low intensity in either frequency of contact (e.g., occurring only on a monthly basis) [137] or scope (e.g., not involving individuals other than the target child) [138]. Further, even targeting early or during periods of rapid growth does not necessarily mean that we can expect significant effects or stability of interventions if we use relatively short-term interventions, even at the correct time period. For children with developmental delays associated with a single specific factor, such as prenatal iodine deficiency, relatively short-duration interventions may be appropriate. However, the situation appears to be quite different for children who are at risk due to exposure to multiple psychosocial and biological risk factors. Particularly for children encountering multiple biological and psychosocial risk factors, there is surprising convergence in the literature on the idea that a three-year intervention period may be necessary to see long-term effects of intervention [131,139].
Even if we utilize interventions of suitable duration and intensity, restricting the interventions just to early development may be problematical, in part because the early developmental period is not the only period of rapid developmental change. We know from the literature that major developmental changes have been noted across cultures both in the five- to seven-year-old period [140] and in adolescence [141]. In addition, even if we get significant gains from early intervention, this does not rule out the likelihood that even greater gains can occur if the interventions are continued over a longer time period. For example, research on the impact of extending pre-school intervention for at-risk inner-city minority children to the early grades clearly showed better academic achievement occurring with the combination of both pre-school and early-grade intervention, rather than either pre-school or early-grade intervention taken in isolation [142].
Further, the fact that we can change an individual’s developmental course by intervention at one particular point in time does not necessarily mean that these gains will be maintained across time. As a general rule, particularly in the case of children exposed to multiple biological and psychosocial risks, it seems clear that there is an increased probability of losing initial benefits unless follow-up interventions are built into the intervention programme. Calling for follow-up interventions immediately raises the question of how much later intervention is needed to maintain the impact of early interventions. Although there is all too little information in this area, one potential clue comes from studies on the concept of reinstatement, the stabilizing effects of periodic partial repetitions of a previous intervention [143]. Although reinstatement phenomena have been most often studied in regard to issues like the persistence of memory [144,145] and relapse into drug addiction [146], the concept of reinstatement may also have implications for understanding how best to maintain the impact of prior psychosocial interventions. Within a reinstatement paradigm, complete re-exposure to a prior intervention experience may not be necessary. Rather, periodic partial re-exposure to key elements of the original intervention experience may be sufficient to maintain the impact of the original intervention.
Human behaviour and development is a highly complex phenomenon that will require more complex intervention strategies than those traditionally utilized. Although more complex intervention strategies can be more costly, it is also important to recognize that complex interventions do not necessarily involve tremendous costs. Further, providing low-cost interventions that have small effect sizes that are not maintained across time may be a far less cost-effective strategy than providing more complex interventions that match the complexity of human developmental variability and that are more likely to be maintained across time and to result in greater effect sizes.
During the writing of this paper, the author was supported by a faculty fellowship for study in a second discipline (nutrition) from Purdue University.
1. Kagitcibasi C. Family and human development across cultures. Mahwah, NJ, USA: Erlbaum, 1996.
2. Masten A, Coatsworth J. The development of competence in favorable and unfavorable environments. Am Psychol 1998;53:205-20.
3. Sinha S. Child rearing practices relevant for the growth of dependency and competence in children. In: Valsiner J, ed. Comparative, cultural and constructivist perspectives. Vol. 3. Norwood, NJ, USA: Ablex, 1995:108-37.
4. Super C, Harkness S. The developmental niche. Int J Behav Dev 1986:9:545-69.
5. Nsamenang A. Human development in cultural context. Newbury Park, Calif, USA: Sage, 1992.
6. Heber R, Dever R. Research on education and habilitation of the mentally retarded. In: Haywood C, ed. Social cultural aspects of mental retardation. New York: Appleton-Century-Crofts, 1970:395-427.
7. Serpell R. Assessment criteria for severe intellectual disability in various cultural settings. Int J Behav Dev 1988;11:117-44.
8. Zigler EN, Berman W. Discerning the future of early childhood intervention. Am Psychol 1983,38:894-906.
9. Werner E. Protective factors in individual resilience. In: Meisels S, Shonkoff J, eds. Handbook of early childhood intervention. Cambridge: Cambridge University Press, 1990:97-116.
10. Zeitlin M. Nutritional resilience in a hostile environment. Nutr Rev 1991;49:259-68.
11. Werner E, Smith R. Vulnerable but invincible. New York: McGraw-Hill, 1982.
12. Radke-Yarrow M, Sherman T. Hard growing. Children who survive. In: Rolf J, Masten A, Cicchetti D, Neuchterline K, Weintrub S, eds. Risk and protective factors in the development of psychopathology. Cambridge: Cambridge University Press, 1990:97-119.
13. Elder G, Van Nguyen T, Caspi A. Linking family hardship to children’s lives. Child Dev 1985;56:361-75.
14. Hoffman L. The influence of the family environment on personality Psychol Bull 1991;110:187-203.
15. Langlois J, Ritter J, Casey R, Sawin D. Infant attractiveness predicts maternal behaviors and attitudes. Dev Psychol 1982;31:464-72.
16. Scheper-Hughes N. Culture, scarcity and maternal thinking. In: Scheper-Hughes N, ed. Child survival. Dordrecht, Netherlands: Reidel, 1987:187-208.
17. Brody G, Stoneman Z. Child competence and developmental goals among rural black families. In: Sigel I, DeLisi A, Goodnow J, eds. Parental belief systems. 2nd ed. Hillsdale, NJ, USA: Erlbaum, 1992:415-31.
18. Neisser U, Boodo G, Bouchard T, Boykin A, Brody N, Ceci S, Halpern D, Loehlin J, Perloff R, Sternberg R, Urbina S. Intelligence: knowns and unknowns. Am Psychol 1996;51:77-101.
19. Halverson C, Kohnstamm D, Martin R. The developing structure of temperament and personality from infancy to childhood. Hillsdale, NJ, USA: Erlbaum, 1994.
20. Rahim S, Cederblad M. Effects of rapid urbanization on child behavior and health in a part of Khartoum, Sudan. J Child Psychol Psychiatry 1984,25:629-41.
21. Vogt L, Jordan C, Tharp R. Explaining school failure, producing school success. Anthrop Educ Q 1987;18:276-86.
22. Diaz-Guerrero R. The development of coping style. Hum Dev 27:320-31.
23. Bolger K, Patterson C, Thompson W, Coopersmidt J. Psychosocial adjustment among children experiencing persistent and intermittent family economic hardship. Child Dev 1995;66:1107-29.
24. Saco-Pollitt C, Pollitt E, Greenfield D. The cumulative deficit hypothesis in the light of cross cultural evidence. Int J Behav Dev 1985;8:75-97.
25. Moffitt T. Adolescent limited and life course persistent anti-social behavior. Psychol Rev 1993;100:674-701.
26. Patterson G, Bank L. Some amplifying mechanisms for pathologic processes in families. In: Gunnar M, Thelan E, eds. Systems and development. Minnesota Symposium on Child Psychology. Vol 22. Hillsdale, NJ, USA: Erlbaum, 1989:167-209.
27. Coie J. Toward a theory of peer rejection. In: Asher S, Coie J, eds. Peer rejection in childhood. Cambridge: Cambridge University Press, 1992:365-401.
28. Wachs TD. Known and potential processes underlying developmental trajectories in childhood and adolescence. Dev Psychol 1996;32:796-801.
29. Dasen P. The cross-cultural study of intelligence. Int J Psychol 1984,19:407-34.
30. Rothbart M. Temperament: a developmental framework. In: Strelau J, Angleitner A, eds. Explorations in temperament. New York: Plenum, 1997:61-74.
31. Levine J, Resnick L, Higgins E. Social foundations of cognition. Annu Rev Psychol 1993;44:585-612.
32. Cohen D, Trayer S. Overlap among domains in conduct disordered and comparison youth. Dev Psychol 1996;32:988-98.
33. Masarro D. Speech perception by ear and eye. Hillsdale, NJ, USA: Erlbaum, 1987.
34. Luthar S, Doernberger C, Zigler E. Resilience is not a uni-dimensional construct. Dev Psychopathol 1993; 5:703-17.
35. Selhub J, Rosenberg I. Folic acid. In: Ziegler E, Filer L, eds. Present knowledge in nutrition. 7th ed. Washington, DC: International Life Sciences Institute Press, 1996:206-19.
36. Stanbury J. Iodine deficiency and the iodine deficiency disorders. In: Ziegler E, Filer L, eds. Present knowledge in nutrition. 7th ed. Washington, DC: International Life Sciences Institute Press, 1996:378-83.
37. Wachs TD. But not sufficient: multiple influences on human development. Washington, DC: American Psychological Association (in press).
38. Plomin R. The role of inheritance in behavior. Science 1990;248:183-8.
39. Bender B, Linden M, Robinson A. Environment and developmental risk in children with sex chromosome abnormalities. J Am Acad Child Adolesc Psychiatry 1987;26:499-503.
40. Carlin M. The improved prognosis in cri-du-chat (5p-) syndrome. In: Fraser W, ed. Key issues in mental retardation research. London: Routledge, 1990:64-73.
41. Plomin R, DeFries J, McClearn G, Rutter M. Behavioral genetics. 3rd ed. New York: Freeman, 1997.
42. Plomin R. Genetics and experience. Thousand Oaks, Calif, USA: Sage, 1994.
43. Finn J. Withdrawing from school. Rev Educ Res 1989; 59:117-42.
44. Gorman K, Pollitt E. School efficiency in rural Guatemala. Int Rev Educ 1992;38:519-34.
45. Ceci S. How much does schooling influence general intelligence and its cognitive components? Dev Psychol 1991;27:703-22.
46. Wachs TD, McCabe G, Yunis F, Kirksey A, Harrison G, Galal 0, Jerome N. Relation of nutritional intake and context to cognitive performance of Egyptian adults. Intelligence 1996;22:129-54.
47. Pollitt E, Gorman K, Engle P, Martorell R, Rivera J. Early supplementary feeding and cognition. Monogr Soc Res Child Dev 1993;58(17):1-99.
48. Werner E. A cross cultural perspective on infancy. J Cross Cult Psychol 1988;19:96-113.
49. Engle P, Zeitlin M, Medrano Y, Garcia L. Growth consequences of low income Nicaraguan mothers’ theories about feeding one year olds. In: Harkness S, Super C, eds. Parents’ cultural belief systems. New York: Guilford, 1996:428-46.
50. Bicego T, Boerma J. Maternal education and child survival. Soc Sci Med 1993:36:1207-27.
51. LeVine R, Miller P, Richman A, LeVine S. Education and mother infant interaction. In: Harkness S, Super C, eds. Parents’ cultural belief systems. New York: Guilford, 1996:254-69.
52. Grantham-McGregor S, Walker S. Health and nutritional determinants of school failure. In: Pan American Health Organization Scientific Monographs 566. Washington, DC: Pan American Health Organization, 1998:82-90.
53. Maughn B. School influences. In: Rutter M, Hay D, eds. Development through life. Oxford: Blackwell, 1994: 134-58.
54. Nettles J, Pleck J. Risk, resilience and development. In: Haggerty R, Sherrod L, Garmazy N, Rutter M, eds. Stress, risk and resilience in children and adolescents. Cambridge: Cambridge University Press, 1994:147-82.
55. Scheerens J, Vermuelen A, Pelgrum W. Generalizability of instructional and school effectiveness indicators across nations. Int J Educ Res 1989:13:789-99.
56. Schoggen P, Shroggen M. Student voluntary participation and school size. J Educ Res 1988:81:288-93.
57. Stromquist N. Determinants of educational participation and achievement of women in the third world. Rev Ed Res 1989:59:143-83.
58. Weinstein C. The classroom as a social context for learning. Annu Rev Psychol 1991:42:493-525.
59. Bronfenbrenner U, Ceci S. Nature-nurture, reconceptionalized in developmental perspective. Psychol Rev 1994:101:568-86.
60. Cadoret R, Troughton E, Mecchant L, Whitters A. Early life psychosocial events and adult affective symptoms. In: Robins L, Rutter M, eds. Straight and devious pathways from childhood to adulthood. Cambridge: Cambridge University Press, 1990:300-13.
61. Casare P, DeVries L, Marlow N. Prenatal and paranatal risk factors for psychosocial development. In: Rutter M, Casear P, eds. Biological risk factors for psychosocial disorders. Cambridge: Cambridge University Press, 1991: 139-74.
62. Kendler K, Eaves L. Models from joint effects of genotype on environment on liability to psychiatric illness. Am J Psychiatry 1986:143:279-99.
63. Wachs TD. Relation of mild to moderate malnutrition to human development. J Nutr Suppl 1992;125: 2245S-54S.
64. Werner E. Overcoming the odds. Dev Behav Pediatr 1994:15:131-6.
65. Grantham-McGregor SM, Chang S, Walker SP. Evaluation of school feeding programs. Some Jamaican examples. Am J Clin Nutr 1998;67:785S-9S.
66. Wachs TD. The nature of nurture. Newbury Park, Calif, USA: Sage, 1992.
67. Bendersky M, Lewis M. Environmental risks, biological risks and developmental outcome. Dev Psychol 1994:30:484-94.
68. Diamond A. The development and neural basis of memory function as indexed by the AB and delayed response tasks in human infants and human monkeys. In: Diamond A, ed. The development and neural basis of higher cognitive functions. Ann NY Acad Sci 1990:608:267-309.
69. Squire L, Zola-Morgan S. The medial temporal lobe memory system. Science 1991:253:1380-6.
70. Vorhees C, Mollnow E. Behavioral teratogenesis. In: Osofsky J, ed. Handbook of infant development. 2nd ed. New York: Wiley, 1987:913-71.
71. Barker D, Gluckman P, Godfrey K, Harding J, Owens J, Robinson J. Fetal nutrition and cardiovascular disease in adult life. Lancet 1993:341:938-41.
72. Wachs TD, Bishry Z, Sobhy A, McCabe G, Shaheen F, Galal 0, eds. Relation of rearing environment to adaptive behavior of Egyptian toddlers. Child Dev 1993:67:586-604.
73. Steinberg L, Lamborn S, Darling N, Mounts N, Dornbusch S. Over time changes in adjustment and competence among adolescents from authoritative, authoritarian, indulgent and neglectful families. Child Dev 1994:65:754-70.
74. Brody G, Stoneman Z, Flor D. Parental religiosity, family processes and youth competence in rural two-parent African American families. Dev Psychol 1996:32:696-706.
75. Gore S, Eckenrode J. Context and process in research on risk and resilience. In: Haggerty R, Sherrod L, Garmazy N, Rutter M, eds. Stress risks and resilience in children and adolescents. Cambridge: Cambridge University Press, 1994:19-63.
76. Leklem J. Vitamin B-6. In: Ziegler E, Filer J, eds. Present knowledge in nutrition. 7th ed. Washington, DC: International Life Sciences Institute Press, 1996:174-83.
77. Pollitt E. A critical view of three decades of research on the effects of chronic energy malnutrition on behavioral development. In: Schürch B, Scrimshaw M, eds. Chronic energy deficiency. Lausanne, Switzerland: International Dietary Energy Consultative Group, 1988: 77-93.
78. Simeon D, Grantham-McGregor S. Nutritional deficiencies and children’s behavior and mental development. Nutr Res Rev 1990:3:1-24.
79. Keusch G. Impact of infection on nutritional status. Am J Clin Nutr 1977:30:1233-5.
80. Ge X, Conger R, Cadoret R, Neiderhiser J, Yates W, Troughton E, Stewart M. The developmental interface between nature and nurture. Dev Psychol 1996;32: 574-89.
81. Quinton D, Rutter M, Gulliver L. Continuities in psychiatric disorders from childhood to adulthood in the children of psychiatric patients. In: Robins L, Rutter M, eds. Straight and devious pathways from childhood to adulthood. Cambridge: Cambridge University Press, 1990:259-78.
82. Coulton C, Korbin J, Su M, Chow J. Community level factors in child maltreatment rates. Child Dev 1995;66: 1262-76.
83. Magnusson D, Bergman L. A pattern approach to the study of pathways from childhood to adulthood. In: Robins L, Rutter M, eds. Straight and devious pathways from childhood to adulthood. Cambridge: Cambridge University Press, 1990:101-15.
84. Bradley R, Whiteside L, Mundfrom D, Casey P, Kelleher K, Pope S. Early indications of resilience and their relation to experiences in the home environments of low birth weight premature children living in poverty. Child Dev 1994;65:346-60.
85. Bornstein M. Sensitive periods in development. Psychol Bull 1989;105:179-97.
86. Georgieff M. Nutritional deficiencies as developmental risk factors. In: Nelson C, ed. Threats to optimal development. Hillsdale, NJ, USA: Erlbaum, 1994:145-59.
87. Goodman R. Developmental disorders and structural brain development. In: Rutter M, Casaer P, eds. Biological risk factors for psychosocial disorders. Cambridge: Cambridge University Press, 1991:20-49.
88. Martorell R, Khan L, Schroader D. Reversibility of stunting. Eur J Clin Nutr 1994;48:S45-S57.
89. Pollitt E, Watkins W, Husaini M. Three month nutritional supplementation in Indonesian infants and toddlers benefits memory function 8 years later. Am J Clin Nutr 1997,66:1357-63.
90. Rutter M. Developmental psychopathology as a research perspective. In: Magnusson D, Casaer P, eds. Longitudinal research on individual development. Cambridge: Cambridge University Press, 1993:127-52.
91. Jacoby E, Cueto S, Pollitt E. Benefits of a school breakfast program among Andean children in Juarez, Peru. Food Nutr Bull 1996;17:54-64.
92. Simeon D, Grantham-McGregor S. Effects of missing breakfast on the cognitive functions of school children of differing nutritional status. Am J Clin Nutr 1989;49: 646-53.
93. Chase-Landsdale P, Lindsay C, Andrew J, Kiernen K. The long term effects of parental divorce on the mental health of young adults. Child Dev 1995;66:1614-34.
94. Levitsky D, Strupp B. Malnutrition and the brain. J Nutr 1995;125:2212S-20S.
95. Lozoff B. Exploratory mechanism for poorer development in iron-deficient anemic infants. In: Pan American Health Organization Scientific Monographs 566. Washington, DC: Pan American Health Organization, 1998:162-78.
96. Rutter M. Statistical and personal interactions. In: Magnusson D, Allen V, eds. Human development. New York: Academic Press, 1983.
97. Thompson W, Grusec W. Studies of early experience. In: Mussen P, ed. Carmichael’s manual of child psychology. New York: Plenum, 1970:565-656.
98. Sroufe L, Egeland B. Illustrations of person-environment interaction from a longitudinal study. In: Wachs TD, Plomin R, eds. Conceptualization and measurement of organism-environment interaction. Washington, DC: American Psychological Association, 1991:68-86.
99. Morrison S, Ames E, Chisholm K. The development of children adopted from Rumanian orphanages. Mer-Palm Q1995;41:411-30.
100. My Lien N, Meyer K, Winick N. Early malnutrition and later adoption. Am J Clin Nutr 1977;30:1734-5.
101. Winick M, Meyer K, Harris R. Malnutrition and environmental enrichment by early adoption. Science 1975;190:1173-5.
102. Hodges J, Tizard B. IQ and behavioral adjustment of ex-institutional adolescents. J Child Psychol Psychiatry 1989;30:53-75.
103. Hodges J, Tizard B. Social and family relationships of ex-institutional adolescents. Child Psychol Psychiatry 1989:30:77-97.
104. Pavenstedt E. The drifters: children of disorganized lower class families. Boston, Mass, USA: Little, Brown, 1967.
105. Devlin B, Daniels M, Roeder K. The heritability of IQ. Nature 1997;388:468-71.
106. Goldberg G, Prentice A. Maternal and fetal determinants of adult diseases. Nutr Rev 1994;52:191-200.
107. Juraska J. The structure of the rat cerebral cortex. In: Kolb B, Tees R, eds. The cerebral cortex of the rat. Cambridge, Mass, USA: MIT Press, 1990:483-505.
108. Karlberg J, Jalill F, Lamb B, Low L, Yeung C. Linear growth retardation in relation to three phases of growth. Eur J Clin Nutr Suppl 1994;48:S25-S34.
109. Schürch B. Malnutrition and behavioral development: the nutrition variable. J Nutr 1995;125:2255S-62S.
110. Colombo M, de la Parra A, Lopez I. Intellectual and physical outcome of children undernourished in early life is influenced by later environmental conditions. Dev Med Child Neurol 1992;34:611-22.
111. Punamaki R. Historical, political and individualistic determinants of coping modes and fears among Palestinian children. Int J Psychol 1988;23:721-39.
112. Patterson J, McCubbin H. The impact of family life events and changes on the health of a chronically ill child. Fam Relat 1983;32:255-64.
113. Golden M. Is complete catch-up possible for stunted malnourished children? Eur I Clin Nutr 1994;48:S58-S71.
114. Gorman K. Malnutrition and cognitive development. J Nutr 1995;125:S2239-44.
115. Sepulveda J, Willett N, Munoz A. Malnutrition and diarrhea. Am J Epidemiol 1988;127:365-76.
116. Strupp B, Levitsky D. Enduring cognitive effects of early malnutrition. J Nutr 1995;125:2221S-32S.
117. Rutter M, Quinton D, Hill J. Adult outcome of institution reared children: males and females compared. In: Robins L, Rutter M, eds. Straight and devious pathways from childhood to adulthood. Cambridge: Cambridge University Press, 1990:135-57.
118. Myers R. The twelve who survived. London: Routledge, 1992.
119. Liddell C, McConville C. Starting at the bottom: towards the development of an indigenous school readiness program for South African children being reared at home. Early Child Dev Care 1994;97:1-15.
120. Wasik B, Ramey C, Bryant D, Sparling J. A longitudinal study of two early intervention strategies. Child Dev 1990;61:1682-96.
121. Lazar I, Darlington R. Lasting effects of early education. Monogr Soc Res Child Dev 1982; 47:no. 2 and 3.
122. Liddell C, Rapodile J, Masilea P. The design and evaluation of a preschool enrichment package for Black South African children in day care. Early Child Dev Care 1991; 66:1-13.
123. Yip R, Dallman R. Iron. In: Ziegler E, Filer L, eds. Present knowledge in nutrition. 7th ed. Washington, DC: International Life Sciences Institute Press, 1996:277-92.
124. Mull D. Traditional perceptions of marasmus in Pakistan. Soc Sci Med 1991;32:175-91.
125. Bronfenbrenner U. Is early education effective? Washington, DC: Department of Health, Education and Welfare, Publication no. OHD76-30025,1974.
126. Williams T. Infant development and supplemental care. Hum Dev 1982;20:1-30.
127. Rosado J, Lopez P, Munoz E, Martinez H, Allen L. Zinc supplementation reduced morbidity, but neither zinc nor iron supplementation affected growth or body composition of Mexican preschoolers. Am J Clin Nutr 1997;65:13-19.
128. Sazawal S, Black R, Bahn M, Jalla S, Bhandari N, Sinah A, Mhaumdar S. Zinc supplementation reduces the incidence of persistent diarrhea and dysentery among low socioeconomic children in India. J Nutr 1996;126: 443-50.
129. Shonkoff J. Health surveillance and the development of children. In: Friedman S, Haywood C, eds. Developmental follow-up. San Diego, Calif, USA: Academic Press, 1994:113-28.
130. Kotchabhakdi N, Winichagoon P, Smitasiri S, Dhanamitta S, Valyasevi A. The integration of psychosocial components in nutrition education in Northeastern Thai villages. Asia Pac J Public Health 1987;1:16-25.
131. Grantham-McGregor S, Powell C, Walker S, Chang S, Fletcher P. The long term follow-up of severely malnourished children who participated in an intervention program. Child Dev 1994;65:428-39.
132. Liddell C, Masilela P. The use of preschool education programs on radio and television by Black South African children. J Broadcast Electr Media 1990;34:85-92.
133. Baydar N, Brooks-Gunn J, Furstenberg F. Early warning signs of functional illiteracy. Child Dev 1993;64:815-29.
134. Cravioto J, De Licardie E. Environmental correlates of severe clinical malnutrition and language development in survivors from kwashiorkor or marasmus. In: Pan American Health Organization. Nutrition, the nervous system and behavior. Scientific Publication no. 251. Washington, DC: Pan American Health Organization, 1972:73-94.
135. Sameroff A. Ecological perspectives on longitudinal follow-up studies. In: Friedman S, Haywood C, eds. Developmental follow-up. San Diego, Calif, USA: Academic Press, 1994:45-66.
136. Idjradinata P, Watkins W, Pollitt E. Adverse effect of iron supplementation on weight gain of iron replete young children. Lancet 1994;343:1252-4.
137. Powell C, Grantham-McGregor S. Home visiting of varying frequency and child development. Pediatrics 1989;84:157-64.
138. Hauser-Cram P, Pierson D, Walker D, Tivan T. Early education in the public schools. San Francisco, Calif, USA: Jossey-Bass, 1991.
139. Super C, Herrera M, Mora J. Long term effects of food supplementation and psychosocial intervention on the physical growth of Colombian infants at risk of malnutrition. Child Dev 1990:61:29-49.
140. Cole M. Culture and development. In: Bornstein M, Lamb M, eds. Developmental psychology. Hillsdale, NJ, USA: Erlbaum, 1992:731-89.
141. Petersen A. Adolescent development. Annu Rev Psychol 1988;39:583-607.
142. Reynolds A. Effects of a preschool plus follow-up intervention for children at risk. Dev Psychol 1994;30: 787-804.
143. Rovee-Colier C. The ontogeny of learning and memory in human infancy. In: Kail R, Spear N, eds. Comparative perspectives on the development of memory. Hillsdale, NJ, USA: Erlbaum, 1984:103-34.
144. Howe M, Courage M, Bryant B. Reinstating preschoolers’ memories. Dev Psychol 1993;29:854-69.
145. McClelland J, McNaughton B, Reilly R. Why there are complementary learning systems in the hippocampus and neocortex. Psychol Rev 1995;102:419-37.
146. Carroll M, Comer S. Animal models of relapse. Exp Clin Psychopharm 1996;4:11-18.
