Date of Submission


Date of Award


Institute Name (Publisher)

Indian Statistical Institute

Document Type

Doctoral Thesis

Degree Name

Doctor of Philosophy

Subject Name

Computer Science


Economics and Planning Unit (EPU-Delhi)


Sen, Arunava (EPU-Delhi; ISI)

Abstract (Summary of the Work)

Standard choice models generally assume that a decision maker (henceforth DM ) chooses an alternative from a set of alternatives. It is often the case that a DM encounters the alternatives in a particular structure. For instance, a shopper has to choose from a set of items displayed on a shelf or a judge has to choose an winner from the contestants who appear one after another in a row. In both these cases, the alternatives appear sequentially to the DM, i.e. the set of alternatives appears in the form of a list. The structure of the set can be more complicated as illustrated in the following example: suppose the DM is purchasing a product online. For every feature of the product the DM gets to choose from a variety of options that appear as hyper-links in the webpage. After choosing an option, a new webpage opens up and provides options for another feature or attribute. Thus the DM has to choose sequentially and the whole set of alternatives appears in the form of a tree in which alternatives labeled at the terminal nodes are ordered, i.e. in the form of an ordered-tree. Some elections may also follow a sequential process which can be modeled aschoosing from an ordered tree. A natural question that arises here is the following: does the structure of the set of alternatives affect the choice or in other words, does it matter how the alternatives are presented to the DM? We can identify several effects which suggest that the choice depends on the structure. We give some examples. When the DM chooses from a list, the first few alternatives may grab attention and become favorite. On the other hand, the DM may be more likely to remember the alternatives that come in the tail of the list. There can be other cases also. For instance, the DM may pay more attention to the alternative which is more distinct than the alternatives that surround it in the list. Similarly, while choosing from an ordered-tree, the DM can be naive, i.e. chooses a particular branch (say, the left-most branch) at each decision node. The DM may also choose a path that is shortest from the initial decision node, because it saves time. A number of experimental and empirical findings indicate that the structure of set of alternatives affects decision. Rubinstein et al. (1996) consider a two-person game where one player hides a treasure in one of four places located in a row and the other player seeks it. They find that both players prefer middle positions to end-points. Attali and Bar-Hillel (2003) find that in an examination with multiple-choice questions, question-setters have tendency to keep the right answer in middle positions and students have tendency to seek it in middle positions. Christenfeld (1995) finds that while purchasing items from a shelf of grocery shop, buyers are naive towards middle positions. Online purchases of items like books, apparel and household items are very common nowadays.


ProQuest Collection ID:

Control Number


Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.