Hierarchical Linear Modeling of Creative Artists' Problem Solving Behaviors*

College art students were videotaped creating original drawings from an array of objects. Judges reliably assessed the creativity of the drawings. Videos of the creation of ten high‐ and ten low‐rated drawings were coded frame‐by‐frame to quantify the extent to which artists engaged in several categories of activities (selecting objects, selecting media, pausing, drawing objects, drawing other visible elements, and drawing imaginary elements) plus reworking the drawing (erasing and revising). Video coding was used to model how the frequency of each measure varied throughout the sessions. Behaviors were analyzed by hierarchical linear modeling, a regression technique that permits individual‐ and group‐level analyses simultaneously. Analyses revealed substantial individual and group differences in the behavioral trajectories, especially in the drawing measures and in reworking the drawing. The results support a positive association between creativity and opportunistic problem solving strategies, and a negative association between creativity and the pre‐planned application of default problem representations.