From exploration to efficiency: Using hidden Markov models to uncover adults’ strategy use on a digital problem-solving task

Adults developing literacy skills often struggle to coordinate tools in digital environments, yet how they navigate digital tasks is not well understood. This study uses PIAAC process data to model navigation on a complex digital problem-solving item. Hidden Markov models identified latent strategy states and transition probabilities by literacy level (developing vs. higher) and by item accuracy. Patterns observed in the higher literacy group included direct, tool-supported, routinized transitions; patterns observed among adults developing their literacy skills included exiting early or displaying diffuse exploration. Among correct solutions, adults with higher literacy showed greater efficiency (e.g., sorting before composing). Among incorrect solutions, adults with higher literacy showed interface uncertainty despite entering specialized states, whereas adults developing literacy skills tended to exit early. Findings clarify how literacy and strategy use intersect in a digital problem-solving context and suggest instruction in adult education that pairs foundational information and communication technology (ICT) practice with planning, tool choice, and verification.