8. Cross-Sectional Analysis

  1. Peter M. Fayers1,2 and
  2. David Machin3,4

Published Online: 8 MAR 2002

DOI: 10.1002/0470846283.ch8

Quality of Life: Assessment, Analysis and Interpretation

Quality of Life: Assessment, Analysis and Interpretation

How to Cite

Fayers, P. M. and Machin, D. (2000) Cross-Sectional Analysis, in Quality of Life: Assessment, Analysis and Interpretation, John Wiley & Sons, Ltd, Chichester, UK. doi: 10.1002/0470846283.ch8

Author Information

  1. 1

    Medical Research Council Clinical Trials Unit, London, UK

  2. 2

    Unit of Applied Clinical Research, Norwegian University of Science and Technology, Trondheim, Norway

  3. 3

    NMRC Clinical Trials & Epidemiology Research Unit, Singapore

  4. 4

    School of Health and Related Research, University of Sheffield, UK

Publication History

  1. Published Online: 8 MAR 2002
  2. Published Print: 18 APR 2000

ISBN Information

Print ISBN: 9780471968610

Online ISBN: 9780470846285

SEARCH

Keywords:

  • nominal data;
  • ordered categorical data;
  • numerical discrete/numerical continuous;
  • categorical data;
  • covariates;
  • variance;
  • normality;
  • ANOVA models

Summary

This chapter describes the types of analysis and graphical methods appropriate to the comparison of groups of patients with QoL assessment at a common time point. The methods described are thus cross-sectional in nature. Some basic ideas for assessing statistical significance when comparing two groups are discussed, such as z and t-tests, and the associated confidence intervals and the extension of these to the comparison of several groups using analysis of variance (ANOVA). These ideas are then generalised to enable differences to be adjusted for covariates, such as baseline patient characteristics (e.g. age and gender), using regression techniques. The relationship between ANOVA and linear regression is illustrated. Modifications to the methods for binary and ordered categorical variables are included, as well as methods for continuous data that do not have the Normal distribution form. Finally, some graphical methods of displaying cross-sectional data are introduced.