| This Ph.D. dissertation is focused on determining quality factors for extra virgin olive oil (EVOO). Since quality is a multidimensional, subjective and variable concept, different approaches and tools are used including descriptive analysis, consumer preference ratings, experts' quality ratings. These findings are further specified in reference to consumer's demographics and behavioral information. Twenty two EVOOs spanning several olive cultivars, countries and prices are selected to cover the spectrum of sensory properties typically encountered in EVOO. The first study discusses the sensory properties of California and imported EVOOs using a generic descriptive panel, and its comparison to the information provided using the attributes measured in the International Olive Council (IOC) certification method. The second study describes how experts perceive quality in extra virgin olive oil. Experts are homogenous in their concept of quality; however they differ in the ratings assigned to each EVOO (ANOVA p<0.05). Positive drivers of quality by experts are ripe and green fruit, nutty, and grassy aromas, bitterness, and pungency.;The third study identifies consumer's preferences and attitudes regarding EVOO, and how hedonic scores are related to quality ratings by experts and descriptive analysis measurements. Cluster analysis reveals segmentation for the EVOO preferences. The three segments agree in the rejection of bitterness and pungency. In general the positive drivers of liking are nutty, tea, green fruit, and green tomato. Some consumers are less sensitive to the presence of defects in EVOO and tend to like defective oils. There are some discrepancies regarding what an expert consider a high quality EVOO versus consumer's preferences.;Finally the last study compares alternative statistical methods---3-PLS, L-PLS, and logistic regression models---versus some of the analyses conducted in Chapter 2, 3 and 4---preference mapping, PLS regression, generalized Procrustes analysis, among others. While the principles of the new methods differ similar results are obtained. The 3-PLS regression model shows that there were agreement among the panelists in the descriptive analysis, at the same time it also identified the drivers of liking for each consumer segment. The L-PLS regression show important demographic characteristics and attitudes related to the overall liking for the products. The logistic regression models identify the main sensory properties associated with the overall liking (acceptance), purchase intention, second consumption and recommendation to friends by EVOO consumers. |
