Elevated corticosterone levels to stress have been found in adult rats exposed prenatally toalcohol, but little is known about the effects of prenatal alcohol exposure on the cortisol responsein humans. To date, one study has found that crack/cocaine was related to depressed newborncortisol levels following a heel prick. In the present study saliva samples were obtained beforeand after a blood draw from 83 inner-city African American 13-month-old infants exposedprenatally to alcohol, cocaine, and other illicit drugs. Post-blood draw cortisol levels did notdiffer from basal levels in many of the infants, confirming recent studies indicating adaptation ofthe adrenocortical response to this type of stress at this age. Maternal depression and emergenceof teeth were positively related to cortisol levels. Alcohol exposure was related to elevated basallevels, cocaine to lower basal levels. As predicted from animal findings, heavy alcohol exposurewas related to elevated poststress cortisol levels.

A group of 164 children from different infant temperament categories were seen at 7 years ofage for a laboratory battery that included behavioral and physiological measurements. The majorresults indicated that children who had been classified as high reactive infants at 4 months of age,compared with infants classified as low reactive, (a) were more vulnerable to the development ofanxious symptoms at age 7 years, (b) were more subdued in their interactions with a femaleexaminer, (c) made fewer errors on a task requiring inhibition of a reflex, and (d) were morereflective. Further, the high reactives who developed anxious symptoms differed from the highreactives without anxious symptoms with respect to fearful behavior in the second year and, atage 7 years, higher diastolic blood pressure, a narrower facial skeleton, and greater magnitude ofcooling of the temperature of the fingertips to cognitive challenge. Finally, variation inmagnitude of interference to fearful or aggressive pictures on a modified Stroop procedure failedto differentiate anxious from nonanxious or high from low reactive children.

During the past 10 years nearly 80 studies on disorganized attachment involving more than6,000 infant–parent dyads have been carried out. The current series of meta-analyses haveestablished the reliability and discriminant validity of disorganized infant attachment. Althoughdisorganized attachment behavior is necessarily difficult to observe and often subtle, manyresearchers have managed to become reliable coders. Furthermore, disorganized attachmentshows modest short- and long-term stability, in particular in middle class environments, and it isnot just a concomitant of constitutional, temperamental, or physical child problems. Thepredictive validity of disorganized attachment is established in terms of problematic stressmanagement, the elevated risk of externalizing problem behavior, and even the tendency ofdisorganized infants to show dissociative behavior later in life. In normal, middle class families,about 15% of the infants develop disorganized attachment behavior. In other social contexts andin clinical groups this percentage may become twice or even three times higher (e.g., in the caseof maltreatment). Although the importance of disorganized attachment for developmentalpsychopathology is evident, the search for the mechanisms leading to disorganization has juststarted. Frightening parental behavior may play an important role but it does not seem to be theonly causal factor involved in the emergence of disorganized attachment.

The purpose of this study was to use the Attachment Q-Set to create categories of attachmentorganization that were conceptually consistent with organizational categories derived from otherattachment assessments, reliable, and valid. We were particularly interested in creating categoriesthat would describe insecure attachment organizations in children from difficult lifecircumstances. Our sample of relationships included 3062 teacher–child relationships.Children ranged in age from toddlers to kindergartners and were diverse in ethnic background.Two percent of the children were diagnosed with severe social and emotional problems andenrolled in a therapeutic preschool. Sixteen percent were enrolled in intervention programs onthe basis of family poverty and another 34% of the children attended subsidized child careprograms for low-income children. All of the children's child–teacher relationshipswere assessed with the Attachment Q-Set (AQS). Independent observers observed or ratedchildren's peer play. Teachers completed behavior problem questionnaires and StudentTeacher Relationships Scales. Using the AQS we created five subscales and six attachmentorganizational categories. These subscales and organizational categories were associated intheoretically meaningful ways with measures of behavior problems, social competence withpeers, and teacher perceptions of child–teacher relationship.

Few studies have examined the association between attachment and psychopathology insamples of clinically diagnosed children or the relation between attachment variables and thecontinuity of clinical symptoms. Clinic-referred preschool boys (n = 80) with earlyonset conduct problems (as indicated by diagnoses of Oppositional Defiant Disorder) werecompared with a case-matched group of 80 nonproblem boys to determine whether observedattachment behaviors could: (a) differentiate the groups, (b) explain concurrent variation inproblem severity within the clinic group, and (c) predict the outcomes of clinic boys 1 and 2years later. Over half of the clinic boys with early onset conduct problems (54%) exhibited aninsecure attachment strategy during reunion, as opposed to 18% of comparison group boys. Allinsecure patterns were overrepresented in the clinic group. Clinic boys were more likely thancomparison boys to engage in provocative behavior when separated from their mothers. Contraryto our hypothesis, attachment variables showed little ability to explain concurrent or futurevariation in the problem severity or diagnostic status of clinic boys. We conclude that observableattachment behaviors exhibited during the preschool years are related to the initial identificationof early onset conduct problems but may have little value by themselves in predicting thesubsequent course and severity of problems in the 2 years following clinic referral.

We tested three competing models regarding the role of deviant friends in the trajectorylinking early disruptiveness with later conduct problems through the use of a preventiveintervention program. The program was implemented during the second and third grade. Onemodel predicted that the program would positively affect later conduct problems by facilitatingnondeviant peer association during early adolescence. The second model predicted a directimpact of the program on later conduct problems through the reduction of early disruptiveness.The third model predicted an interaction between postintervention disruptiveness and associationwith less deviant friends. The results showed that the program's effects on later conductproblems were mediated by the reduction in disruptiveness and by the association with lessdeviant friends. However, the positive effect of associating with less deviant friends depended onwhether children's disruptiveness had been reduced or not by their participation in theprogram, thus supporting the third model. We recommend using intervention studies to testdevelopmental models.

A three-wave longitudinal study design with two 6-month intervals was used to examine thestability and change in Internalizing, Externalizing, and Total Problem behavior among childrenand adolescents referred to outpatient mental health services. Our results indicated highstabilities for parent ratings of child psychopathology across a 1-year interval. Additionally, wefound decreases in the level of problem behavior. Interindividual differences in change werefound for Total Problems and Externalizing behavior but not for Internalizing. While both thechild's temperament and intelligence level and family relations were related to theinitial level of parent-rated problem behavior, only intermediary stressful life events had aninfluence on the rate of change of child psychopathology.

Most previous research has assumed that adolescent alcohol use and problem use represent acontinuum and are influenced by the same psychosocial factors, with problem use representingmore severe psychosocial impairment. The current study evaluated this assumption by identifyingthe correlates of adolescent alcohol use and those of problem use. Using a community sample ofadolescent children of alcoholics (COAs) and a demographically matched comparison group(non-COAs), a typology of adolescent alcohol use was created, and alcohol use groups werecompared on variables chosen from nine psychosocial domains. The correlates of problemalcohol use were different from those of moderate use. Problem use was associated withfundamental family disruptions and poor psychological functioning. In contrast, the determinantsof moderate alcohol use reflected unconventionality and socialization specific to alcohol. Fewpsychosocial variables distinguished abstainers from light drinkers. Intervention andmethodological implications of these findings are discussed.

Although a majority of adolescent mothers are graduating from high school, the processesthat enhance the educational attainment of adolescent mothers are not well understood. With asample of 93 African American and Puerto Rican adolescent mothers, we assessed the effects ofmaterial support from family (i.e., child care assistance from grandmother and residence withgrandmother) and emotional support from family over and above pre- and postpregnancy riskfactors (i.e., maternal age at first birth, delayed grade placement before pregnancy, ethnicity,depressive symptoms, stressful life events, and repeat pregnancy) during the first year postpartumon educational attainment at 6 years postpartum. Significant contributors to the explainedvariance in educational attainment included: delayed grade placement before pregnancy, maternalage at first birth, depressive symptoms, emotional support from family, and residence withgrandmother. Unexpectedly, higher perceived emotional support from family and living withgrandmother predicted lower educational attainment. Post hoc analysis of qualitative datasuggested reasons for these latter findings and point to the need to reconceptualize and broadenthe elements of social support that constitute protective factors for adolescent mothers.

The term self-organization has at least two meanings for developmentalpsychopathologists: the spontaneous emergence of order in developing systems and theconsolidation of that part of the person called the self. The first meaning is shared across thesciences at large, and it denotes a general model of change and growth in complex systems. Wereview principles of self-organization from the natural sciences and discuss their application tothe study of human development. We then suggest possible sources and consequences ofterminological confusion in developmental psychology and offer some potential resolutions. The“self” in self-organization seems a special case of this confusion, and it can beaddressed by modeling the self as both autonomous agent and emergent form. We outlinepreliminary directions for this modeling and then go on to a more general discussion of howprinciples of self-organization can help bring developmental psychopathology closer to itsgoals.