This study examined the structure of self-reported problem behaviors in Mainland Chinese children. Third grade (n ϭ 3011) and fifth grade (n ϭ 3708) students completed a 20-item Problem Behavior Frequency Checklist (PBFC). Findings suggested that among the four factors studied (physical aggression, substance use, delinquency, and self-centered behaviors), self-centered behaviors and physical aggression were reported most frequently by the Chinese elementary school children. Furthermore, boys had higher prevalence rates of problem behaviors than girls. Developmental differences were also present. Confirmatory factor analyses supported a model that included specific factors related to substance use, physical aggression, delinquency, and self-centered behaviors. Therefore, the single factor conceptualization in which different types of problem behaviors are hypothesized to reflect a single underlying common factor may not hold for this population. © 2003 Wiley Periodicals, Inc.For past two decades, behavioral dysfunction in children and adolescents has been an important field for research. Numerous studies have found that "problem" behaviors, such as substance use, aggression, delinquency, and early sexuality, are positively correlated (e.g., Donovan & Jessor, 1985;Donovan, Jessor, & Costa, 1988;Gillmore, Hawkins, Catalano, Day, Moore, & Abbott, 1991;Grube, Morgan, & Seff, 1989;Vulcano, Barnes, & Langstaff, 1990). According to the problem behavior theory (Jessor & Jessor, 1977), a framework for understanding the relationship between problem behaviors, the strong positive association between problem behaviors can be attributed to a single common factor, namely, unconventionality (Donovan & Jessor, 1985;Donovan, Jessor, & Costa, 1988;Donovan, Jessor, & Costa, 1991). However, several studies have questioned the single factor model of problem behavior (e.g., Farrell, Kung, White, & Valois, 2000;Gillmore et al., 1991;Grube & Morgan, 1990;Osgood, Johnston, O'Malley, & Bachman, 1988;McGee & Newcomb, 1992;Resnicow, Ross-Graddy, & Vaughan, 1995;Williams, Ayers, Abbott, Hawkins, & Catalano, 1996). For example, Farrell and his colleagues, using confirmatory factor analysis, identified distinct domains of problem behaviors (aggression, drug use, and delinquency) that could only partially be explained by an underlying factor. Moreover, the work of Achenbach and his colleagues (e.g., Achenbach & Edelbrock, 1987;Achenbach, 1991;McConaughy & Skiba, 1993) have been noted as an important contribution to research of problem behavior. Their work has led to the distinction between internalizing problem behaviors (i.e., withdrawn, somatic complaints, anxiety and depression) and externalizing problem behaviors (i.e., defiance, impulsivity, disruptiveness, aggression, antisocial features, and over-activity).The statistical method used for data analyses provides further challenges to the single factor model of problem behavior. For example, in studies where confirmatory factor analysis (CFA), one type of structural equation modeling (SEM), ...