Are smarter people more procedure invariant? An online experiment investigating the effects of cognitive ability on preference reversal
Notes
My undergraduate dissertation. It involved a huge amount of research, as well as building and running an online experiment and analysing the results. I hugely enjoyed putting it together, and it has ended up being one of the most rewarding things I have ever done. It was later described by my supervisor as "a genuine addition to knowledge" and given a mark of 91% — the highest amongst my year group of 300 at the University of Nottingham. I even won a £1000 prize for it.
Abstract
In the preference reversal literature, models of stochastic choice and error are increasingly endorsed as explanations of the phenomenon. In this paper, I document an online experiment examining whether there exists the negative link between cognitive ability and observation of preference reversal that an error-led approach may imply. I find that cognitive ability has no effect on whether subjects exhibit PR or not, but that when subjects do so, the extent of reversal is strongly and negatively linked to cognitive ability. I find that part of this effect is caused by increased simple error making amongst lower cognitive ability subjects, but argue that further research is needed to establish the full nature of the association, incorporating more detailed notions of error and cognitive ability.
Keywords: preference reversal, cognitive ability, errors, decision making, risky choice
JEL classification: C90, D01, D81
Downloads
Full text
Appendices
Data are anonymous and presented in a random order.
- Core dataset: Excel
- Full dataset: Excel
- Instructions/screenshots/parameters: PDF
- Beginners' guide and explanation: Blog post
License
Licensed under Creative Commons Attibution-NonCommercial 3.0.