1996 2004 1995 2002 1993 2006 2000 1993 1999 1975 2000 1985 1993 1996 1996 2000 1985 1993 1995 The criteria on which we based our selection of the various predictors and our hypotheses regarding their joint contribution to the development of externalizing problems – led by various interrelated theoretical models – are described per domain in the following paragraphs. Parent-child interaction and parent-child attachment 1975 parent-child interactions 2002 2006 1987 1997 1996 2003 1996 infant-parent attachment 1978 1986 1990 1978 1990 1990 1999 1985 1990 2001 1996 1985 1990 disorganized 1999 1999 1993 2001 In view of the above, we expected infant attachment insecurity and particularly attachment disorganization to be related to elevated levels of externalizing behavior at age 5. Considering that attachment patterns have been found to reflect the history of parent-child interactions, we expected the association between the early parent-child interaction and later externalizing problems to be at least partially mediated by the quality of the infant-parent attachment. Child characteristics 2004 2006 1996 1996 1997 1993 1998 1998 2001 2006 2005 2004 2002 1993 2006 2001 2003 Parental characteristics parental personality 1994 1994 2003 1980 2002a Contextual characteristics 1984 1993 stressful life events 2000 partner support 1991 1996 socio-economic status 2006 1996 1983 1996 In sum, the aim of the present study was to predict the occurrence of externalizing behavior problems in a sample of 5-year-old children on the basis of various parental, child, dyadic, and contextual factors assessed at 15 and 28 months and at age 5. We expected the following specific factors to independently or interactively contribute to the development of externalizing problems: (a) a low quality of parent-child interaction at 15 and 28 months; (b) disorganized parent-infant attachment; (c) child temperament (i.e., high anger proneness and low fearfulness) at both 15 and 28 months and particularly in interaction with low quality parenting; (d) low levels of child cognitive ability at 15 months; (e) low parental ego-resiliency; (f) high incidence of stressful life events between 15 months and 5 years; (g) lack of partner support; and (h) low SES. Furthermore, the contribution of parental ego-resiliency to child externalizing behavior was expected to be mediated by the quality of parent-child interaction, and the contribution of the early parent-child interaction by the quality of infant-parent attachment. Method Participants M SD n SD M SD M SD 1 t t p Procedure 15-month data 2002a 1978 2002a 2002b 28-month assessment age-5 assessment Instruments and measures Quality of parent-child interaction (at 15 and 28 months) 1985 1985 2000a effective guidance negative interaction Quality of parent-child attachment (at 15 months) 1978 1984 1979 2000 1978 1990 Child temperament (at 15 and 28 months) 1994 Anger proneness Social fear Cognitive ability (at 15 months) 1969 1983 M SD Parental ego-resiliency (at 15 months) 1961 1991 Stressful life events (at 28 months and 5 years) 1978 1972 2001 1996 Partner support (at 15 months) 1997 2002a 2006 Socioeconomic status (SES at 15 months) 1996 1996 Externalizing problems as rated by parents and teachers (5 years) 1991a 1996 α 1991b α Results 2003 Preliminary analyses 1 1999 M 1 Table 1 Intercorrelations, means and standard deviations for the study variables  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 15-month predictors                    1. Effective guidance –                   2. Negative interaction .00 –                  a ** ** –                 a ** * ** –                a −.05 .11 ** −.12 –               a * ** ** −.16 −.12 –              7. Child social fear ** -.03 * −.16 −.06 −.05 –             8. Child anger proneness .04 .02 .02 .06 −.00 −.09 ** –            9. Child cognitive ability ** −.09 .09 −.07 .05 −.08 .18 .12 –           10. Child sex .10 −.01 .10 .03 −.15 −.05 .08 .07 .15 –          11. Parental ego-resiliency * −.07 .17 .07 ** −.09 −.00 −.07 * −.05 –         12. Partner support .16 −.03 .16 .02 * −.07 −.06 −.01 .12 .05 .17 –        13. SES .05 −.17 −.02 .02 .10 −.07 .11 −.02 .14 −.02 .17 −.07 –       28-month predictors                    14. Effective guidance ** −.06 * * .15 −.18 * .10 ** .04 .09 −.07 .19 –      15. Negative interaction −.09 ** ** .11 −.05 ** −.11 −.02 −.07 −.08 ** −.10 ** .00 –     16. Child social fear * .14 −.04 −.15 .10 .13 ** .16 .03 .15 −.15 −.17 .05 .17 .02 –    17. Child anger proneness −.01 −.05 −.02 .15 .01 −.13 * ** .09 .07 −.02 −.15 .08 −.05 −.15 .14 –   5-year predictor                    18. Life events 15 mo-5 yr .02 * −.06 −.15 .10 .16 −.04 .12 −.12 −.03 * ** −.10 .10 .19 .16 * –  5-year outcome                    19. Externalizing behavior * ** ** −.02 .05 ** −.18 −.02 −.09 ** ** ** −.18 −.17 ** −.11 .11 * – M b b 0.64 0.14 0.08 0.14 3.68 3.38 103.90 0.47 0.44 8.32 c b b 3.38 3.61 1.48 c SD 1.00 1.00 0.48 0.35 0.27 0.34 0.85 0.72 16.93 0.50 0.19 2.05 2.81 1.00 1.00 0.80 0.72 1.31 1.73 Note. N a b c * p ** p Table 2 Hierarchical regression results for predicting age 5 externalizing scores (Total Model)   Hierarchical regression results Block  B SE B β 2 1 a 1.88 0.42 ** 22.2  Negative interaction 28 months 0.65 0.15 ** 10.7  b 0.61 0.26 * 4.3  Anger proneness 28 months 0.47 0.18 ** c  Partner support −0.14 0.06 * 2.6 2 Interaction terms      Effective guidance × Anger proneness 28 months −0.33 0.14 ** 3.3    R 2    F ** a D D b c * p ** p SD N n T 1991a SD N n T 1991b M SD 1996 M SD 1997 t p t p t t p p Total Externalizing t M boys SD M girls SD t p Predicting age 5 externalizing scores Hierarchical regression analysis 1991 2005 2 Fig. 1 Child externalizing behavior scores in relation to effective guidance for children high and low on anger-proneness  SD SD 1991 1 B t p B t ns. Mediational analyses 1986 1 effective guidance β p β ns z p negative interactions β p β p z p 1 β p z p β ns) Fig. 2 * p ** p  Path analysis 2003 χ 2 χ 2 df p 2 2 Discussion 2000 1985 1993 1996 2002a not 1996 2000 2003 2006 infancy toddlerhood 1997 1998 2005 Our results on the interaction between temperament and parenting in the development of externalizing problems extend those of earlier studies by showing that a specific temperament by parenting interaction may work particularly in a specific developmental phase. That the interaction between anger proneness and effective guidance proved to predict externalizing problems particularly for toddlers makes sense when interpreted from a developmental psychopathology perspective: establishing autonomy and learning to comply with social rules and expectations are major developmental tasks for toddlers that challenge the parents’ ability to provide effective guidance by imposing structure and setting limits while at the same time remaining emotionally supportive of the child. As a result of the child’s striving for autonomy, the second year is marked by parent-toddler conflicts. Toddlers that are prone to show angry behaviors in such discord situations are particularly in need of effective guidance in order to prevent their lack of emotional and behavioral self-control to further escalate and evolve into externalizing behavior problems. 1993 2001 1996 2001 1999 beyond 1985 The results of this study may have clinical and policy implications. Many studies have shown that children born in disadvantaged environments are at risk for developing externalizing problems later in life. The present study showed that, also for children living in presumably lower-risk families, certain characteristics in infancy and toddlerhood indicate an increased risk of externalizing problems at preschool age. It should be kept in mind, however, that our sample may have been more at risk than originally thought, because a relatively large proportion of children turned out to have externalizing behavior scores above the subclinical cutoff. An important practical implication of our findings is that one of the most powerful predictors identified in the present study, i.e., negative parent-child interactions, is relatively easy to observe already in infancy and proves to be highly stable, which makes it an interesting candidate for inclusion in early community mental health screening procedures. And the specific interaction of the other parent-child interaction factor (i.e., effective guidance) with temperamental anger proneness in toddlerhood in predicting later externalizing problems also suggests that paying more attention to early parent-child interactions may be useful in early detection of children at risk and in preventive intervention programs for parents in community mental health care. But more research is needed, of course, to further explore these possible applications. 1 primary 2006 In sum, the longitudinal model our analyses generated provides more insight into the complex interplay among parental, child, dyadic and contextual characteristics that together shape the development of externalizing behavior from age 15 months onwards. Future research should reveal whether the predictive power of the model can be improved, for instance by including earlier measurements and by taking into account the children’s early experiences with the other parent and with their caregivers and peers in child-care facilities.