In this study a utility maximizing model is developed which accommodates changing states of information. Rational consumer choices in one state of information can lead to realizing different levels of utility than anticipated. Differences between these levels of utility suggest a measure for the value of information. This framework is applied to estimating the potential cost of possible inaccuracies in the EPA fuel-economy ratings. Survey data collected from new car buyers then are used to infer the magnitude of the actual costs that may be caused by the present EPA information.