Structural work on the nature of child temperament has lagged far behind parallel research with adults. Further, the structure of extant parent-report questionnaires has been examined only at the higher-order level, using small samples of children. This dissertation addresses this important gap in three studies, examining the lower-and higher-order structure of two widely used parent-reports of child temperament (Studies 1 and 2) and an observational battery of temperament (Study 3). The first study examined the lower-and higher-order structure Fewer than half of the lower-order structures resembled original CBQ and TMCQ scales.Higher-order EFAs indicated that a four-factor structure consisting of Sensation Seeking, Disinhibition/Anger, Low Negative Affect (NA)/Soothability, and Smiling/Approach was the best fit for preschoolers, and that a three-factor structure consisting of Impulsivity/Anger, NA, and Social Dominance best fit the TMCQ data. The higher-order models obtained were only modestly similar to the models proposed by their developers. In the third study, we extended the scarce extant literature on the structure of observed temperament. Using exploratory structural equation modeling, we derived a four-factor model (Positive Affect/Activity, Impulsivity/Anger, Surgency, and Dysphoria) of preschool-aged temperament. These factors showed meaningful concurrent associations with children's internalizing and externalizing symptoms, providing further support for the use of observational measures. Findings from these studies contribute to the development of a unified taxonomy of child temperament/personality.