Use this URL to cite or link to this record in EThOS:
Title: Response models and efficient designs for change-over experiments with treatment carryover
Author: Ferris, Steven
Awarding Body: University of Edinburgh
Current Institution: University of Edinburgh
Date of Award: 1999
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Humans are used as recording instruments in many areas of scientific experimentation. However, their responses are susceptible to bias due to the context in which the sensory stimuli are presented. One recognised source of bias is carryover, i.e. the effect of the previous stimuli on the current judgement. It is therefore important to take account of carryover effects in both the design and analysis of the experiments in order to obtain precise and bias free estimates of experimental treatment effects. In this study we investigate carryover in two areas: sensory profiling of food products and the assessment of crop disease severity. A series of experiments are designed, conducted and analysed for both applications, in order to ascertain the form, frequency and magnitude of carryover. Alternatives to the standard additive carryover model are proposed for the sensory profiling responses. The proposed model has carryover effects which are proportional to direct treatment effects. In visual assessment carryover is found to depend on whether the previous stimulus is higher than the current stimulus and an appropriate model is developed to describe this relationship. Results for optimal and efficient change-over designs for estimating direct treatment effects in the presence of carryover, in addition to repeat treatment effects, are derived for the proportional carryover model analytically. Balanced uniform designs with or without a circular pre-period for specified design parameters are determined to be optimal within their respective classes of competing designs. The search for optimal and efficient change-over designs is extended to all possible designs using a computer search algorithm. However, the relative efficiency of designs is shown to depend on the value of the proportional scalar linking carryover effects to direct treatment effects, and knowledge of this parameter will influence the optimal design.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available