Application and Analysis of a Cross-Over Design

Clive J. Lawrence, Shahid Kamal, Sandra Lang

Abstract


In this paper we consider the application and analysis of a cross-over design in a medical problem. Some statistical detail is given of the linear model used and its analysis. This design is in the class of repeated measurement analyses which can, for the medical research worker, be a source of some difficulty. Its application to a study of the constituents of breast-milk, and their possible variation between hand and pump expression is examined. Data for this analysis comes from a study that has been carried out in Exeter, UK, in which information was collected on a number of different aspects of infant feeding. Results for sodium, potassium, and other constituents are reported, together with some discussion of the various practical difficulties that can arise in the organisation of studies of this kind. Comments are included concerning the extension from two to three-period designs, and in the testing of significance.

Full Text:

PDF


DOI: http://dx.doi.org/10.18187/pjsor.v1i1.110

Refbacks

  • There are currently no refbacks.




Copyright (c)

Title

Application and Analysis of a Cross-Over Design

Keywords

-

Description

In this paper we consider the application and analysis of a cross-over design in a medical problem. Some statistical detail is given of the linear model used and its analysis. This design is in the class of repeated measurement analyses which can, for the medical research worker, be a source of some difficulty. Its application to a study of the constituents of breast-milk, and their possible variation between hand and pump expression is examined. Data for this analysis comes from a study that has been carried out in Exeter, UK, in which information was collected on a number of different aspects of infant feeding. Results for sodium, potassium, and other constituents are reported, together with some discussion of the various practical difficulties that can arise in the organisation of studies of this kind. Comments are included concerning the extension from two to three-period designs, and in the testing of significance.

Date

2005-07-01

Identifier


Source

Pakistan Journal of Statistics and Operation Research; Vol 1. No. 1, July 2005



Print ISSN: 1816-2711 | Electronic ISSN: 2220-5810