This embedding effect can arise because the respondents give a value that expresses moral satisfaction of contributing to a good cause rather than the value of the specific good. Another possible explanation for the embedding effect is that the respondents mentally value a different good than was intended by the researcher, i.e. the model they use to value the good include benefits and perhaps other goods, which differs from the researchers mental model. Hypothetical bias arises when hypothetical WTP differs from real WTP. When respondents are unfamiliar with the good to value they have no real value for the item. There is therefore a risk that the respondent just makes up values when answering the questionnaire. Both embedding and hypothetical bias are possible to eliminate if the researcher is able to present a hypothetical situation to the respondent that reflects a reliable market situation. Most of the CV research today therefore concerns questions on how to create such a situation by focusing on the study design.
The other approach to estimate individual preferences is to use actual market behavior. Individuals" preferences for a non-marketed good in an RP study are, as mentioned above, estimated by observing the behavior of the individuals on an existing market. The advantage with the RP method in comparison to the SP methods to estimate individuals" preferences for safety is that real market behavior is observed. The disadvantage is that safety per se is often embedded among other characteristics of a good or service purchased on the market and it is therefore difficult to isolate the variable of interest, e.g. the safety component of a car. The method used to estimate preferences and derive implicit values by market behavior is called The Hedonic Pricing Method (HP) and was developed by Lancaster, Griliches and Rosen. The idea behind the method is that goods embody a bundle of characteristics, and that the price.
