Empirical studies of the role of non-cognitive skills in driving economic behavior often rely heavily on the assumption that these skills are stable over the relevant time frame. We analyze the change in a specific non-cognitive skill, i.e. locus of control, in order to directly assess the validity of this assumption. We find that short- and medium-run changes in locus of control are rather modest on average, are concentrated among the young or very old, do not appear to be related to the demographic, labor market, and health events that individuals experience, and are unlikely to be economically meaningful. Still, there is no evidence that locus of control is truly time-invariant implying that the use of lagged measures results in an errors-in-variables problem that could downward bias the estimated wage return to locus of control by as much as 50 percent. Those researchers wishing to analyze the economic consequences of non-cognitive skills should consider (i) restricting their analysis to the working-age population for whom there is little evidence of systematic change in skill levels and (ii) accounting for error in the skill measures they employ.