The course of development is shaped by numerous interlocking factors from across domains, some of which are tied to specific characteristics of individual children and others that result from external circumstances. Forging an understanding of the components of learning and behavior requires acknowledging the presence and influence of these moving pieces. This dissertation proposed that observing children’s play and examining relations among skill assessments are effective methods for uncovering the mechanistic processes underlying developmental change. Its studies specifically targeted the interplay of spatial-cognitive, motor, language, and visual processing abilities. Chapters 2 and 3 documented age differences in how 12- to 48-month-old children approached playing with toys that involved spatial reasoning, such as shape sorters and puzzles. The studies in these chapters also indicated that children’s ability to successfully solve the spatial reasoning tasks posed by these toys varied according to their perceptual and tactile features.Chapter 4 situated the Beery Developmental Test of Visuomotor Integration (DTVMI; Beery, 2004), a standardized assessment of children’s visual perception and motor coordination, within a greater cross-domain context by relating it to a gradient of spatial-cognitive assessments and a visual processing assessment. Four- and five-year-old children’s performance on the Beery DTVMI significantly predicted their performance on a mental rotation task and a spatial language assessment, providing justification for its use as a measure of spatial cognition in addition to motor and visual domains. Overall, these studies provide a framework for thinking about multidimensionality in early childhood research in an observable and testable way. By examining how children’s spatial play behaviors evolve over age and vary according to toy design, we gain a deeper appreciation of the intermediary steps in their cognitive and motor planning development. Likewise, contextualizing existing skill assessments such as the Beery DTVMI in a new light facilitates discovery of connections among skills that may have otherwise been overlooked. Embracing the complexities in this type of multilayered data helps to approximate the complexity of the developmental processes themselves.