Contingent valuation is a non-market valuation technique which elicits preference data from participants by asking them to value a change in the provision of a specific good or service, contingent on the specifications outlined in a hypothetical market scenario. A commonly observed feature of CV results is a significant and pervasive disparity that develops between willingness to accept and willingness to pay measures of value. While these two measures should theoretically produce the same value estimate for a particular good, with the exception of a small difference due to the income effect, this is not the case in a majority of the experimental CV literature. This WTA-WTP disparity is the focus of this thesis’ investigation, which aims to offer a more accurate understanding of the phenomenon. This thesis provides a detailed review of past research experiments investigating the WTA-WTP gap, identifying two main alternate explanations for the disparity: one which explains the discrepancy as resulting from flaws in the methodological design of the CV experiments (weak experimental design) and one which suggests the gap is caused the fact that people place a higher value on a good they own than an identical good they do not own (endowment effect theory). To assess the legitimacy of each of these two explanations, this thesis presents an experimental investigation into the WTA-WTP gap, where a basic CV survey is design and then used to elicit preference data from participants for organic cotton. The experimental design includes six CV surveys, all of which are fundamentally identical except for small specific alterations, which will allow valuation results to be compared across survey groups in an attempt to isolate the effect that the specified survey design features have on valuation estimates. Two of the surveys collect WTP and WTA data from participants under a binding condition where the average valuation stated by the group would determine a binding monetary outcome for all of the participants. Two further surveys collect WTP and WTA data from participants where no binding monetary outcome is specified (i.e. purely hypothetical), and the final two treatment groups are asked to estimate the WTP and WTA of the binding groups, rather than provide their own personal value estimates. The core comparisons possible between these survey designs include: assessing whether a WTA-WTP gap is observed even when controlling for features of weak experimental design, assessing how the hypothetical nature of a CV experiment impacts on the valuation results, and whether participants are able to provide an unbiased estimate of others’ preferences. Data was collected from 178 participants with between 27 and 31 respondents involved in each of the six survey groups. The data was then analysed using SPSS to test whether there were significant differences between the valuation estimates collected from the different participant groups. The results of the experiment found that the WTA-WTP gap is caused by the endowment effect rather than weak experimental design, that hypothetical and binding treatments do not differ significantly in terms of valuation estimates, and that participants are able to provide unbiased estimates of others’ preferences, so long as they are not first asked to state their own preferences.