Flexible use of tactics in Sudoku

When solving problems people flexibly apply different tactics. Here, we use Sudoku to study this flexibility. In a think-aloud study participants use at least two tactics, a cell-based and a number-based tactic, and have personal preferences for one or the other. Response times in two follow-up experiments indicate that participants can be biased towards either tactic by task instructions and task requirements. We argue that previous research often used biasing task designs and therefore underestimated participants' flexibility and overestimated the importance of a problem's complexity. Furthermore, our experiments demonstrate that if a tactic does not lead to a solution, participants are able to switch to the other. We model each tactic and we show that only by incorporating switching we can fit the data.