Individual Open-Ended Problem Solving and Creativity

Abstract

Complex real-world problems are often ill-structured and open-ended, with no single correct solution and many possible ways to organize the available information. Despite the prevalence of such problems, most experimental research on problem solving has focused on well-defined tasks with predefined solutions, while studies of ill-structured problems have largely relied on qualitative observational methods. This thesis extends a controlled experimental framework for studying open-ended problem solving, originally developed and tested in a group context by Alattas (2023), to the level of the individual solver.

Sixty participants each completed three categorization tasks in which they organized 16 randomly selected pictures into four categories of four pictures each. Task open-endedness was manipulated by varying the goal structure and participants' beliefs about the solution space across three conditions: an Expert condition, in which participants were told to find the single best solution as identified by a panel of experts; a Good condition, in which multiple acceptable solutions were described as available; and a Story condition, in which participants were asked to form narrative-based categories. All participants completed all three conditions with different stimulus sets, using a within-subjects design.

Quantitative analyses examined five outcome domains: task difficulty, solution variability, path dependency, association strength, and search behavior. Open-endedness significantly increased solution variability and path dependency, reduced association strength, and led to more direct search paths with fewer direction reversals. However, the effect on task difficulty diverged from the group-level pattern: Expert condition produced the highest difficulty across behavioral measures, while Good and Story were comparable, rather than the graded Expert > Good > Story ordering observed in groups. Post-experiment interviews were analyzed for 30 participants using an inductive thematic approach. Three main themes emerged: interpreting the pictures before categorizing, strategies for building categories under constraint, and using narratives and creativity to reach a solution. These themes illuminate the processes behind the quantitative patterns, revealing how participants interpreted ambiguous stimuli, managed constraints, and sometimes used stories to connect the pictures across all three conditions.

Together, the quantitative and qualitative findings show that the effects of open-endedness on solution variability, association strength, and search behavior observed by Alattas (2023) in groups extend to individual solvers, whereas the relationship between open-endedness and task difficulty shifts at the individual level. The study contributes an experimental method for studying individual open-ended problem solving, provides early evidence on how the shift from group to individual processing changes the relationship between open-endedness and difficulty, and raises questions about the relationship between open-endedness and creativity that invite further investigation.