Measurement Error or Individual Variation

Among other things, I will focus on some studies where preferences are measured at the individual level. Since we cannot look into the head of an individual, we must use indirect measurements. We therefore use a model. The model describes the assumed structure and enables a discussion of estimation methods, as well as a discussion of what is estimated and, not least, allows the model to discuss the interpretation of estimates. Whether estimates should be interpreted at the individual level or at the population level, and the difference will often reflect the purpose of the analysis.<br/>I will give some examples of how it is important in relation to the interpretation of an analysis how estimates are interpreted.