This thesis uses the Decision by Sampling model as a basis for examining effects of rank order encoding in value judgement and preferential choice. A range of experiments are reported, and these employ a variety of methodologies including behavioural paradigms, eye tracking and functional MRI. The results show that when there are a relatively small number of values used during an experiment, participants encode utility based upon the rank order of a potential outcome within these values. By introducing different decision contexts where the experienced values have a positive or negative skew, an individual’s utility curve can be made concave and risk averse or convex and risk seeking. These different utility curves can be produced within the same individual and same task simply by providing a contextual cue for each trial. Two fMRI experiments demonstrate the neural systems underlying this phenomenon. The results show that all regions of the reward network encode reward as a function of the reward’s rank order within the current context. No region of the brain was found to encode a reward’s absolute financial value. Other experiments investigated choice and valuation in more complex decision environments. It was found that when the number of experienced values is significantly larger than working memory capacity DbS is a relatively poor predictor of behaviour. The Weighted ADDitive rule proved to be more accurate throughout. However, in multi-attribute choice experiments where one attribute had a manipulated distribution, individuals use and weighting of the attribute value was determined by rank order rather than its numerical value. The specific characteristics of this were found to be incompatible with an exemplar based model of recall and binary comparison to specific items. It was instead found to be compatible with non-exemplar, fuzzy trace theories of decision making which are based upon estimates of the distribution. Eye tracking during multi-attribute choice additionally shows that participants begin to attend more to their preferred choice as they near the point at which they respond. However, they do not attend more to the attributes which they weight more highly in their choices, questioning the validity of previous eye-tracking findings.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:686873 |
| Date | January 2014 |
| Creators | Mullett, Timothy L. |
| Publisher | University of Nottingham |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://eprints.nottingham.ac.uk/14293/ |
Page generated in 0.0025 seconds