On the persistence of strategic sophistication

We examine whether the ‘Level-k’ model of strategic behavior generates reliable cross-game testable predictions at the level of the individual player. Subjects’ observed levels are fairly consistent within one family of similar games, but within another family of games there is virtually no cross-game correlation. Moreover, the relative ranking of subjects’ levels is not consistent within the second family of games. Direct measures of strategic intelligence are generally not correlated with observed levels of reasoning in either family. Our results suggest that the Level-k model is just one of many heuristics that may be triggered in some strategic settings, but not in others.