A stated preference analysis comparing the Analytical Hierarchy Process versus Choice Experiments Zein Kallas a,⇑ , Fatima Lambarraa b , José Maria Gil a a Center for Agro-food Economy and Development and Polytechnic University of Catalonia, (CREDA-UPC-IRTA), Edifici ESAB, C/ Esteve Terrades, 8, 08860 Casteldefells, Barcelona, Spain b Courant Research Center ‘‘Poverty, Equity and Growth”, University of Göttingen, Platz der Göttinger Sieben 3, 37073 Göttingen, Germany article info Article history: Received 16 September 2009 Received in revised form 16 September 2010 Accepted 24 September 2010 Available online 5 November 2010 Keywords: Analytical Hierarchy Process Choice Experiments Consumer preferences abstract Our paper compares individual preferences for attributes and levels of an agro-food product using two different methodologies. A key question is whether different results are obtained when asking consumers what they ‘‘prefer” versus what they would ‘‘buy”. Within this context, we compare empirical results obtained from Choice Experiments (CE) versus the Analytical Hierarchy Process (AHP) due to their capac- ity to simulate the ‘‘purchasing” and ‘‘theoretical” stated preferences, respectively. We utilize a survey of restaurateurs’ stated preferences to include rabbit meat in their menus in Catalonia (Spain) in this explor- atory study. Results demonstrate similarities and differences in the stated preferences for attributes and attribute levels with a 55.6% coincidence in their ranking between methods. Utility variation between levels in both approaches follows a similar shape for two of the three analyzed attributes. Ó 2010 Elsevier Ltd. All rights reserved. 1. Introduction and objectives Several alternatives are available when analyzing stated prefer- ences. The Choice Experiment (CE) is one of the most frequently used in the exploration of individual preferences (Alfens, 2004; Burton & Pearse, 2002; Burton, Rigby, Young, & James, 2001; Carlsson, Frykblom, & Lagerkvist, 2007). This method has a demon- strated capacity to analyze preferences for ‘‘complex goods”, i.e., goods that comprise several parts or attributes, such as food prod- ucts. The Analytical Hierarchy Process (AHP) has also been used to assess individual preferences in a hierarchical structure (Parra, Calatrava, & De Haro, 2005; Ramanathan & Ganesh, 1994; Scholl, Manthey, Helm, & Steiner, 2005; Scholz & Decker, 2007; Sedef, Bülent, & Canan, 2007). The latter technique allows for the search of relative importance placed on product attributes and attribute levels of the analyzed complex goods. In this context, both tech- niques are able to construct a ranking structure of the product attributes and attribute levels offering the opportunity to compare ranked results between the techniques. The CE belongs to the stated preference method which is based on the creation of a hypothetical market for the analyzed goods and services. This approach tries to simulate the ‘‘purchasing sta- ted preference” of individuals since one of the attributes is usually the price of the product. Individuals are asked which product they would buy from a set of competitive products at different prices. The decision to make a trade-off between attributes is based on an individual cost-benefit analysis or cost-satisfaction analysis. However, in the AHP technique, product attributes and attribute levels are evaluated in a direct pairwise comparison within a struc- tured hierarchy. Price is not included in the comparison since our study tries to assess the relative importance of non-monetary attri- butes. Furthermore, it does not make sense to realize a pairwise comparison for the different price levels. In this context, the AHP, in studying individual preferences, tries to identify the ‘‘theoretical stated preference” of products’ attributes and levels independent of the purchasing decision. Several studies have compared individual preferences and cus- tomer decisions using AHP and Conjoint Analysis (CA) 1 using a case-ranking method as a means of comparison (Perini, Ricca, & Susi, 2009). CA is closer than AHP to CE (Malvinas, Mangkoesubroto, Sury- adi, & Yudhistira, 2005; Meißner, Scholz, & Decker, 2008; Scholz, Meißner, & Wagner, 2005), however they also differ in two impor- tant ways. First, CE offer two or more ‘‘choice sets” to respondents from which they are asked to choose the one they most prefer, while in CA respondents express their degree of preference for each prod- uct (profile or cards). Second, CE was created to overcome several critical assumptions inherent in the CA design that could lead to incorrect predictions. CE seems to be more realistic and natural for 0950-3293/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodqual.2010.09.010 ⇑ Corresponding author. Tel.: +(34) 93 55 21213; fax: +(34) 93 55 21121. E-mail addresses: Zein.kallas@upc.edu (Z. Kallas), Fatima.lambarraa@upc.edu (F. Lambarraa), Chema.Gil@upc.edu (J.M. Gil). 1 CA tries to determine how people value different attributes of an individual product. It determines what combination of attributes is most influential on respondent choice. Several cards representing the products are shown to individuals asking them to make their preferences. As a result, an implicit valuation (utilities or part-worths) of product attributes are determined. Food Quality and Preference 22 (2011) 181–192 Contents lists available at ScienceDirect Food Quality and Preference journal homepage: www.elsevier.com/locate/foodqual