Based on a synthesis of the literature and observations of young children over 2 years, a framework for assessing probabilistic thinking was formulated, refined, and validated. For each of four major constructs incorporated into this framework--sample space, probability of an event, probability comparisons, and conditional probability--four different levels of thinking were established that reflect a continuum from subjective to numerical reasoning. The framework was validated through data obtained from 24 children in grades 1 through 3 who served as case studies. Results suggest that, while the framework produces a unified picture of children's thinking in probability, there is "static" in the system which generates inconsistencies among construct levels. The framework has implications for curriculum development and assessment. Contains 18 references. (Author)