Analysis of Repeated Measures Data With Illustrations

Garven, Frances (1986) Analysis of Repeated Measures Data With Illustrations. MSc(R) thesis, University of Glasgow.

Abstract

Repeated Measures data arises when a response variable is measured on the same experimental units on two or more occasions. Due to the dependence between observations on the same unit, special statistical methods may be required. This thesis contains a review of experimental designs, models and methods of analysis which may be appropriate when handling data of this type. These methods are illustrated using three practical problems having different types of repeated measures structure. Chapter one examines the structure of repeated measures data. Balanced and unbalanced designs of varying complexity are described with several examples. Split-plot designs and designs resulting in growth curve data are identified as special cases. Chapter two contains some models which may be applied in the analysis of repeated measures data. The univariate and multivariate analysis of variance, the general growth curve and the two stage random effects models are outlined for the analysis of balanced repeated measures data. A generalisation of the general growth curve model is outlined for the analysis of unbalanced data and we note that the two stage random effects model may also be used. Chapter three .reviews some of the approaches and methods which may be used in the analysis of balanced and unbalanced repeated measures data, covering those models contained in chapter two. The methodology discussed includes univariate analysis of variance, approximate and conservative univariate tests, multivariate analysis of variance and some special techniques for the analysis of growth curves. These techniques involve the modelling of the data and the application of either multivariate analysis of variance or covariance. Chapter four Illustrates some of the models and methods discussed in earlier chapters using three practical problems. The first problem entails the analysis of a balanced two factor repeated measures design with two trial factors. The second involves the analysis of a balanced three factor design with two grouping factors and one trial factor. Finally the third problem is an example of an extremely unstructured set of growth curve data where the underlying problem is discrimination.

• Item Type: Thesis (MSc(R))
• Qualification Level: Masters
• Additional Information: Adviser: Ian Ford
• Keywords: Statistics
• Date of Award: 1986
• Depositing User: Enlighten Team
• Unique ID: glathesis:1986-76420
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 19 Nov 2019 14:32
• Last Modified: 19 Nov 2019 14:32
• URI: https://theses.gla.ac.uk/id/eprint/76420