Developmental psychopathology adherents have long theorized that clinical patterns can be conceptualized as deviations from normative patterns (Cicchetti & Richters, 1997
). Furthermore, clinical syndromes are increasingly viewed through a dimensional lens (Hudziak, Achenbach, Althoff, & Pine, 2007
). These perspectives are well-developed theoretically, but their empirical evidence base is less well developed. We tested a dimensional model of Temper Loss
in preschoolers utilizing rigorous psychometric methods applied to a novel, developmentally sensitive, parent-completed questionnaire. Findings hold promise for generating empirically derived parameters for ‘when to worry,’ even during a developmental period which has long challenged nosologic systems. First, we found that temper tantrums occur occasionally in most preschoolers, but only approximately 10% of children exhibit them daily. Second, we psychometrically validated our theory that Temper Loss
falls along a dimensional continuum ranging from normative misbehaviors to problem indicators. Third, we demonstrated associations between Temper Loss
and other clinical problems. These correlational data are but a first glimpse, however, as they are based solely on parent report in the absence of validated indicators of clinical significance and/or impairment. One of the most common conundrums of preschool psychopathology is the murky boundary between normative misbehavior and disruptive behavior. The recent validation of developmentally sensitive diagnostic methods, which rely on observation in standardized contexts and/or parent report, provides techniques for clinical assessment within research contexts (Egger et al., 2006
; Wakschlag et al., 2008
). However, the absence of a standard metric for referral contributes to a double-edged sword of minimizing parental concern and underidentification versus potential overidentification and psychotropic overtreatment (Wakschlag & Danis, 2009
; Zito et al., 2007
). If clinically validated, our findings suggest that two key features of tantrums may be useful for screening in primary care settings.
First, consistent with previous studies, nearly all preschoolers (83.7%) tantrum sometimes, but having daily tantrums is not typical
(Bhatia et al., 1990
; Osterman & Bjorkqvist, 2010
). Our findings suggest that this is true for younger preschoolers, for girls and boys, and for children across varying levels of contextual risk. This provides empirical evidence that inquiring about the frequency of tantrums may yield important information that can help guide clinical concern. Second, quality, as well as frequency, contributes to the severity continuum. Both normative misbehaviors and facets of anger regulation have mild, commonly occurring, as well as severe and more rarely occurring, manifestations. Severe temper loss may manifest as daily tantrums of any form or qualitatively more severe behaviors exhibited less than daily, including tantrums lasting more than 5 minutes, being aggressive during a tantrum, having a tantrum with nonparental adults, and having a tantrum ‘out of the blue.’ Prior work in both clinical (Belden et al., 2008
) and community (Green et al., 2011
; Osterman & Bjorkqvist, 2010
) samples also suggests these features are developmentally meaningful indicators of concern.
The importance of an evidence base for determining the feasibility and clinical utility of taking developmental and sociocultural differences into account has been noted (Frick & Nigg, 2011
; Moffitt et al., 2008
). In the MAPS Study, we present preliminary support for consistent patterns of atypicality across sociodemographic categories. For example, although older preschoolers are less likely to tantrum than younger preschoolers, daily temper loss is not normative at any age. The same is true for girls and boys, poor and nonpoor children, and across racial/ethnic groups. Future research should include fine-grained analyses within subgroups to elucidate variations in thresholds and optimize item parameters.
was associated with parent-reported problems of aggression, hyperactivity, anxiety, and depression. While it is essential to extend these findings with assessments beyond parent-report, this suggests that temper regulation problems may serve as a common psychopathological substrate, as previously demonstrated with older youth (Stringaris & Goodman, 2009
). The strong association between Aggression
and Temper Loss
may be at least partially explained by shared method variance, as both were assessed with the MAP-DB. The MAP-DB Temper Loss
construct also centers solely on angry affect and behavior. Incorporation of developmentally specified expressions of a broader range of negative affect may enhance the MAP-DB Temper Loss
scale’s clinical salience for mood problems. For example, Potegal and colleagues have proposed an ‘anger-distress’ model, which incorporates sadness as a tantrum feature (Green et al., 2011
The dimensionality of temper loss processes may have important implications for clinical conceptualizations and elucidation of mechanisms (Helzer et al., 2008
). Moving beyond a dichotomous characterization to a more nuanced spectrum of temper loss provides a framework for identifying emergent problems and targeting young children at-risk due to family history of psychopathology and/or environmental adversity. It also delineates a range of behaviors that might ultimately serve as intervention targets. Consistent with the NIMH Research Domain Criteria initiative (Insel et al., 2010
), viewing temper dysregulation as a spectrum also holds promise for linkage to neural circuitry and genetic mechanisms in a more fine-grained manner than current categorical classifications allow.
The limitations of this work must be noted. First, these developmentally defined patterns of atypicality are merely suggestive when obtained from a single parental report. To establish clinical validity, MAP-DB Temper Loss must be linked to established measures of disruptive, mood and anxiety disorders and impairment, cross-sectionally, and longitudinally. We are currently collecting data to validate the MAP-DB against multimethod, multiinformant measures, including the PAPA, DB-DOS, teacher questionnaires, and neurocognitive measures. Second, some children with undiagnosed autism spectrum disorders may have been included in analyses, potentially inflating temper tantrum rates. Third, our findings are limited to early childhood. A basic tenet of this study is that developmental specification will enhance characterization of phenotypic heterogeneity and accuracy of identification. Testing life span continuities and incremental utility of developmentally specified dimensional models of temper loss is critical. Fourth, the recall window for Temper Loss frequency in the MAP-DB is limited to the past month to ensure that behaviors were not fleeting. Whereas DSM-IV requires 6-month duration for ODD symptoms, duration thresholds for preschoolers have not been established empirically.