In psychology as well as behavioral economics, it is well established that our choices and judgments are not just a function of the available options, but also of the context surrounding them. Several models have been brought forward to explain these context effects. We use the decision by sampling model (DbS; Stewart, Chater, & Brown, 2006) and investigate possible mechanisms that might lead to the relativity of judgment and choice. Stewart, Reimers and Harris (2015) demonstrated that shapes of utility and probability weighting functions could be manipulated by adjusting the distributions of outcomes and probabilities on offer. Chapter 2 reports a multi-level replication where we find that these effects are robust, but that DbS is unlikely to be the (sole) explanation for its origins. We conclude that problems with revealing utility functions from expected utility fits may be responsible for biasing the shapes of utility functions. Chapter 3 shows that reduced working memory capacity, as manipulated by cognitive load, does not reduce the effects found in Chapter 2. This further points away from a DbS explanation of the above findings. In Chapter 3, we also find that cognitive load has no impact on risk aversion, but find that choice consistency is reduced when working memory capacity is reduced, which also challenges the prominent dual process theories. Still, the question where the differences in preferential functions come from, remains unexplained. Chapter 4 reviews over 20 behavioral as well as neurophysiological studies showing that even if the rank effects are an artefact of the estimation procedure, this does not question the many findings that support a model encompassing a rank-dependent evaluation of alternatives. In Chapter 5, we test this hypothesis in a new design, where a monetary outcome is evaluated in the light of another foregone outcome with a history of other foregone outcomes. In contrast to our hypothesis, we find no evidence for a rank-dependent evaluation here. In Chapter 6, we investigate how comparison processes can lead to the mutable-zero effect. In the mutable-zero effect, participants prefer an outcome entailing a “pay zero” or “lose zero” attribute over an outcome entailing a “receive zero” or “win zero” attribute. We find that the only comparisons that are made with “pay zero” are other payments and that the only comparisons that are made with “receive zero” are other receipts. This process leads to “pay zero” comparing favorably and “receive zero” comparing unfavorably, which in turn leads to “pay zero” options being preferred over “receive zero” options. Given the findings in Chapter 2 and 3 are robust, they point to a general problem with estimating preferential functions using models like expected utility theory or prospect theory. Chapter 3 and 5 were a first attempt at testing a DbS-predicted mechanism: In contrast to our predictions, cognitive load did not decrease context sensitivity. Instead, we observed a slight increase in random choices. Finally, in exploring the mutable zero effect, we added to evidence that comparisons spontaneously happen only within gains, or within losses, but not across gains and losses.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:720458 |
| Date | January 2016 |
| Creators | Canic, Emina |
| Publisher | University of Warwick |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://wrap.warwick.ac.uk/90155/ |
Page generated in 0.002 seconds