The study used computational models of planning to compare with real human behavior. Participants quickly judged which subgoals were both useful and manageable, and they preferred options that lowered expected mental effort later in the task as well as up front. This suggests people carry a sense of future cognitive cost when planning—an internal budgeting of attention and problem solving that goes beyond the next move.
That internal budgeting matters for more than block towers. If we can map how people trade off immediate progress against future effort, we can design tools, instructions, or interfaces that fit natural human planning and broaden who succeeds at hands-on tasks. Follow the full article to see how these findings connect to learning, accessibility, and designing environments that expand human potential.
Abstract
From building a new piece of furniture to replacing a lightbulb, people must often figure out how to assemble an object from its parts. Although these physical assembly problems take on many different forms, they also pose common challenges. Chief among these is the question of how to break a complex problem down into subproblems that are easier to solve. What principles determine why some strategies for decomposing a problem are favored over others? Here, we investigate the decisions that people make when considering different visual subgoals in the context of attempting to build a series of virtual block towers. We hypothesized that people favor subgoals achieving a balance between how much progress the subgoals would help achieve toward the final goal and how effortful they would be to solve. We tested this hypothesis by defining several computational models of planning and subgoal selection, then evaluating how well these models predicted human planning and subgoal selection behavior on the same problems. Our results suggest that participants rapidly differentiated the computational costs of otherwise similarly ambitious subgoals, and used these judgments to drive subgoal selection. Moreover, our findings are consistent with the possibility that participants were not only sensitive to the immediate computational costs associated with solving the very next subgoal, but also future costs that might be incurred when attempting the rest of the problem. Taken together, these results contribute to our understanding of how humans make efficient use of cognitive resources to solve complex, grounded planning problems.