Tailoring value elicitation to decision makers' numeracy and fluency: Expressing value judgments in numbers or words Barbara Fasolo a,n , Carlos A. Bana e Costa b a Department of Management, London School of Economics and Political Science, London WC2A 2AE, United Kingdom b CEG-IST, Centre for Management Studies of Instituto Superior Técnico, Universidade de Lisboa, Avenida Rovisco Pais, 1049-001 Lisbon, Portugal article info Article history: Received 13 September 2011 Accepted 26 September 2013 Available online 8 October 2013 Keywords: Multicriteria Value elicitation Direct rating MACBETH abstract In organizational settings, options evaluation requires managers to express value judgments on multiple criteria. This research investigates the influence of decision makers' numeracy (ability to use appropriate numerical principles) and fluency (ability to express oneself in words) on their subjective experience of value elicitation as supported by two different techniques: direct rating and MACBETH. The former asks for value judgments to be expressed numerically, the latter non-numerically. The results of our experiment indicate that the two techniques are not psychologically equivalent: decision makers with higher numeracy express values more easily when assisted by the numerical technique whereas decision makers with higher fluency find value elicitation easier with the non-numerical technique. These findings highlight the importance of tailoring value elicitation to decision makers' numeracy and fluency. Implications for decision scientists and analysts are discussed. & 2013 Elsevier Ltd. All rights reserved. 1. Introduction Interest in multi-criteria decision analysis increased as the sphere of application of quantitative management science moved from operational decision making situations, for which a more-or- less well-defined single objective function could be identified with little controversy (e.g., maximize profit), to more complex levels of managerial planning and decision making, which are naturally multi- dimensional problems (in p. 30 of [1]). This has also been recognized in engineering [2]. In organizations, the pressure to make defensible value judgments heightens the importance of a consistent approach to elicitation of preferences [3]. Indeed, managers continuously face the task of expressing, and justifying, judgments about the relative attractiveness or value of several options at the level of each criterion individually. For unaided individuals, this process is challenging and sometimes even arbitrary [4]. To assist them, decision analysts often use multicriteria value methods, which are extensively reviewed in the decision analysis literature [3,5–8]. Goodwin and Wright provide a simple and intuitive example of a managerial office location problem to illustrate the usefulness of different techniques to elicit value judgments (from p. 33, on Chs. 3 and 4 of [7]). Experimental research in the management science literature has shown that these techniques to elicit value judgments appeal differently to decision makers (see, for example, [9,10]). This notion has been flagged by Larichev and Brown [11] who suggested (but did not experimentally assess) that preference for verbal versus numerical aiding techniques is a matter of decision analysts' habit and expertise, as well as decision makers' education. Here, we conduct a behavioral experiment to examine the extent to which decision makers' numeracy and verbal fluency impact their perception and preferences for two different value- elicitation techniques, one numerical and one non-numerical. In our experiment we use a laptop choice problem to compare the numerically based direct rating technique used in the Simple multi-attribute rating technique, or SMART approach [12,13] and the non-numerical Measuring Attractiveness by a Categorical- Based Evaluation Technique (MACBETH) [14–16]. These two have been widely applied in public and private managerial and engi- neering settings (see examples [17–22]) and complex policy frameworks [23–27]. Moreover, they are of particular interest because on the one hand they are “technically equivalent”: they are found on the same principles of value-difference measurement [8], both aim at scoring the options on an interval scale of measurement and for this purpose both require comparison of intervals of value during the elicitation process. Interval value scales are quantitative representations of preferences and reflect not only the order of attractiveness of the options, but also differences in their relative attractiveness, or in other words, the strength of the decision maker's preferences for one option over another. On the other hand, the questioning procedures of the two techniques differ: using SMART the analyst asks decision makers Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/omega Omega 0305-0483/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.omega.2013.09.006 n Corresponding author. Tel.: þ44 20 7955 7617; fax: þ44 20 7955 6885. E-mail addresses: b.fasolo@lse.ac.uk (B. Fasolo), carlosbana@tecnico.ulisboa.pt (C.A. Bana e Costa). Omega 44 (2014) 83–90