That's the title of a paper I co-authored with Glynn Tonsor and Ted Schroeder, which is forthcoming in the European Review of Agricultural Economics. The abstract:
In the past two decades, there has been an explosion of studies eliciting consumer willingness-to-pay for food attributes; however, this work has largely refrained from drawing a distinction between preferences for health, safety and quality on the one hand and consumers' subjective beliefs that the products studied possess these attributes, on the other. Using data from three experimental studies, along with structural economic models, we show that controlling for subjective beliefs can substantively alter the interpretation of results and the ultimate implications derived from a study. The results suggest the need to measure subjective beliefs in studies of consumer choice and to utilise the measures when making policy and marketing recommendations.
We show applications related to tenderness, added growth hormones in beef, and country of origin labeling. Here are a couple excerpts:
The reason why the conventional ‘reduced form’ model yields a potentially misleading result is that it does not take into account the fact that most people believe that the generic steak is safe. The reason the premium for natural over generic was so low in the ‘reduced form’ model was not because people did not care about safety but rather because they, on average, believed the health risks from growth hormones and antibiotics in the generic steak were low.
and
the results reveal that, at the mean beliefs, consumers are WTP a premium of only about USD 1.68 . . . for a US origin steak relative to the ‘weighted average origin’ steak. The reason why the value is so low is that most people believe the unlabelled steak is highly likely to come from US origin.
The estimates allow us to make interesting calculations like:
of the total WTP premium for guaranteed tender steak, 46 per cent is due to perceived value of added tenderness; the remaining 54 per cent is due to other factors. A similar computation reveals that of the total WTP premium for natural steak over the generic steak, only 38 per cent is due to perceived added healthiness or no hormone use; the remaining 62 per cent is due to other factors.
and
The implication is that when a product has a mixed-origin label, people are
apparently pessimistic, believing the joint-labelled product to have a much
higher likelihood of coming from the less-preferred origin.