The Impact of Salient Missing Information and Evaluation Mode in Multidimensional Choices

Missing information complicates consumer evaluation of choice alternatives. Such missingness is ubiquitous on the Internet and particularly in relation to comparison websites. Products are often described by a listed set of attributes with their values or levels either listed or not. The listing of attributes regardless of whether level information is available makes missing information especially salient to the consumer.  At the same time, important differences have been demonstrated between joint evaluation situations, where consumers are exposed to multiple alternatives simultaneously and evaluate the alternatives comparatively, and the contrasting separate evaluation situations, where consumers evaluate alternatives in isolation. Thus, this study investigates the impacts of salient missing attribute level information on consumer decision making under single and joint evaluation modes for multidimensional choices. There is a large body of research focusing on how consumers react to missing information (e.g., Johnson and Levin 1985, Jaccard and Wood 1988). This literature focuses on what inferential processes consumers use to fill in missing information for important aspects of the product, in order to make a choice. However, whether a consumer recognises that information is missing and how they process missing information is context specific (Kardes, Posavac, and Cronley 2004). Research is sparse in measuring the specific impact of missing information on choice (Kivetz and Simonson 2000). Previous research has investigated the impact of missing information has focused on changes in aggregate preference, with the exception of Islam, Louviere and Burke (2007), who demonstrated the impact of missing information on both the systematic and random components of utility. Their results suggest that not taking choice variability into account can bias the results, leading to contradictory findings (Islam, Louviere, and Burke 2007). Several studies document systematic changes in preference for alternatives dependent on whether they are evaluated jointly or separately (Hsee and Leclerc 1998; Hsee et al. 1999). Most recently, this finding has been extended to demonstrate the effect of evaluation mode on both aggregate preference and correlational structures of preference in multidimensional choices (Wallin and Coote 2014).Whether a choice context is single versus joint has been found to impact consumer sensitivity to missing information. Specifically, consumers are better able to detect the absence of relevant information and are more sensitive to missing information in a comparative context (Sanbonmatsu et al. 1997). The main goal of this research is to test whether there are systematic differences between the impact of missing information between single and joint decision contexts for multidimensional choices. A further contribution is to explore these effects at the attribute level, demonstrating missing information’s impacts on both aggregate preferences and the way in which consumer’s trade-off attributes (correlational structure of preferences).  The motivation is to extend researchers understanding of how missing information impacts consumer decision making, as well to provide practitioners with guidance on the potential impacts of incomplete sources of product information. A discrete choice experiment (DCE) was developed for each task (single evaluation mode and joint evaluation mode) in the context of purchasing a Smartphone. Each Smartphone handset was characterised by eight attributes (Capacity, Operating System, Camera, Video, Battery, Bluetooth, Key Pad and Price).  The eight attributes were chosen because of their significant impact on preference for Smartphones in previous studies. Respondents for the DCEs were 160 student-consumers choosing to buy or not buy Smartphone handsets in the single evaluation tasks and choosing among alternative from sets of Smartphone handsets (or not buy) in the joint evaluation tasks. An availability design was used to systematically vary the presence/absence and attribute level information presented in each set of choices. An availability design uses a balanced incomplete block design (BIBD) to designate the presence/absence of attribute level information and an orthogonal main effects plan (OMEP) to determine the attribute level information when present (Louviere, Hensher, and Swait, 2000). An OMEP was generated for the 2 5 factorial. To create the second alternative (for the joint evaluation task) a systematic set of level changes was applied to create additional sets of profiles. The design resulted in 8 scenarios (with corresponding pair for joint evaluation task) for each of the seven presence/absence conditions according to the BIBD. A control group was included in the experiment, using a 2 8 factorial to create 12 scenarios with full information for all profiles. The evaluation tasks were counterbalanced. A model catalogue is advanced with varying degrees of latent structure. The most fundamental of the models (McFadden’s conditional logit) explores differences in aggregate preferences due to variation in missing attribute information. Subsequent models add latent structure using the structural choice modelling (SCM) framework. SCM enables the simultaneous estimation of the tasks so that the within-subjects approach is reflected in the modelling and allows the specification of latent factors to reflect the underlying preference structure (Rungie, Coote, and Louviere 2011, 2012). A factor structure reflecting the two evaluation modes is developed to test whether there is a systematic difference in the impact of missing information between single and joint evaluations. Factor analytic models are also used to retrieve consumer’s aggregate preferences and understand the ways in which consumers are trading-off between attributes. The next stage of the analysis will demonstrate the impact of salient missing information on the latent structure of preferences across evaluation modes. In addition, tests will demonstrate that evaluation mode impacts on the evaluability of attribute information will also impact the extent to which missing attribute information is impacting choice. Preliminary results suggest there are systematic and meaningful differences in the impact of missing information across single versus joint evaluation modes, beyond differences in aggregate preferences established by prior research. The findings will have implications for several literatures, including the marketing literature on information processing. References Hsee, C. K., and Leclerc, F. (1998). Will products look more attractive when presented separately or together?. Journal of Consumer Research , 25(2), 175-186. Hsee, C.K., Loewenstein, G.F., Blount, S., and Bazerman, M.H. (1999). Preference reversals between joint and separate evaluations of options: A review and theoretical analysis. Psychological Bulletin , 125(5), 576-590. Islam, T., Louviere, J.J., and Burke, P.F. (2007). Modeling the effects of including/excluding attributes in choice experiments on systematic and random components. International Journal of Research in Marketing , 24, 289-300. Jaccard, J., and Wood, G. (1988). The effects of incomplete information on the formation of attitudes toward behavioural alternatives. Journal of Personality and Social Psychology , 54(4), 580-591. Johnson, R.D., and Levin, I.P. (1985). More than meets the eye: The effect of missing information on purchase evaluations. Journal of Consumer Psychol ogy, 14, 230-256. Kardes, F.R., Posavac, S.S., and Cronley, M.L. (2004). Consumer inference: A review of processes, bases, and judgement contexts . Journal of Consumer Psychology , 14(3), 230-256. Kivetz, R., and Simonson, I. (2000). The effects of incomplete information on consumer choice . Journal of Marketing Research , 37(4), 427-448. Louviere, J.J., Hensher D.A., and Swait, J.D. (2000), Stated Choice Methods: Analysis and Applications . Cambridge: Cambridge University Press. Rungie, C.M., Coote, L.V., and Louviere, J.J. (2011). Structural choice modelling: Theory and applications to combining choice experiments. Journal of Choice Modelling , 4(3), 1–29. Rungie, C.M., Coote, L.V., and Louviere, J.J. (2012). Latent variables in discrete choice experiments. Journal of Choice Modelling , 5(3), 145-156. Sanbonmatsu, D. M., Kardes, F. R., Posavac, S. S., and Houghton, D. C. (1997). Contextual influences on judgment based on limited information. Organizational Behavior and Human Decision Processes , 69(3), 251-264. Wallin, A.C. and Coote, L.V. (2014). An extension and test of the evaluability thesis. Working Paper.