When survey-based, self-reported measures of satisfaction are used as dependent variables in a regression framework, the estimates may be biased and inconsistent. We adopt the measurement error perspective introduced by Bertrand and Mullainathan [2001], and propose an individual fixed-effects technique to tackle some of the consistency issues deriving from the potential correlation between the explanatory variables and the biases in the reported level of satisfaction. We apply the fixed-effects method to the analysis of the correlation between reported levels of satisfaction derived from the consumption of services of general interests collected in the Eurobarometer special surveys, and measures of the regulatory and competitive environment collected in the REGREF database