Economists are still searching for methods to reduce/eliminate Hypothetical Bias (HB). Different methods have been previously applied some with success and others without. In this study, we aimed to further test the cognitive dissonance approach (CD) through a learning design method to estimate the WTP for five beef attributes: Non-quality, Kentucky Proud, Appalachian, Grass Fed and a mix of 25% Non-quality and 75% Kentucky Proud, using a one and one half bounded model. To test the CD, 881 participants from the University of Kentucky, were randomly assigned to a real/hypothetical market for a battery recycling project at first and afterwards to a hypothetical market for beef. For the battery recycling, participants were asked to donate $1, $2 or $3. For the beef market, participants were randomly assigned to a $4 or $6 for the non-quality attribute patty. Participants assigned to a $4 were afterwards randomly assigned to a $4.5, $5, $5.5 or $6 and the ones assigned to a $6 were afterwards assigned to a $6.5, $7, $7.5 and $8 for the other attribute patties. From this study, we found that the learning design was effective in reducing the cognitive dissonance or conflicts between what consumers say and their actions.