Use this URL to cite or link to this record in EThOS:
Title: Modelling latent preferences in a system where choice outcomes are influenced by capacity constraints on some alternatives : a case study of school choice in Northern Ireland
Author: Spollen, M.
ISNI:       0000 0004 5367 3807
Awarding Body: Queen's University Belfast
Current Institution: Queen's University Belfast
Date of Award: 2014
Availability of Full Text:
Full text unavailable from EThOS.
Please contact the current institution’s library for further details.
This thesis develops the theory of latent preference (LP) discrete choice analysis required to facilitate robust analysis of latent preference data. Latency in this context arises when characteristics of the choice system reduce the probability that individuals can realise utility-maximizing alternatives. Such a situation arises when one or more of the available alternatives in the choice set is subject to an externally-imposed supply constraint. When such constraint is binding for any alternative, such that the aggregate expressed demand for that alternative is greater than available supply, then some proportion of all decision-makers who prefer the constrained alternative must select a less preferred alternative instead - one that they would have rejected in the absence of the constraints. In these circumstances, data on choices observed ex post will not reflect the underlying - or latent - preferences to an extent that is proportional to the imbalance between supply and demand at the level of the alternatives. This will be true even where supply in aggregate across all alternatives is greater than aggregate demand. Three new latent preference discrete choice models (LP-DCMs) are developed and applied to the case study of latent preferences for post-primary school in Northern Ireland. These models are referred to as the LP linear, LP non-linear and LP hybrid models. A fourth method, using Monte Carlo simulation, is outlined as a direction for future work. The results reveal an improved fit to the case study data compared to results from existing models, with fitted parameter estimates on key variables observed to change significantly and in a plausible manner. It is hoped that the new models will benefit researchers faced with analysing LP data - such as enrolment data from public schools, health care and housing programs - and offer a viable alternative to recourse to stated preference surveys.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available