Statistical models that analyse time to recurrence of bladder cancer

Abstract

Conventional regression analysis can be utilised to represent the relationship between duration and explanatory variables. However, researchers are aware of the prejudice inference on the response variables committed by the violation of the normality assumption. Thus, conventional regression analysis is rarely used with duration since researchers often give greater importance to the distributional characteristics of duration. Cross-sectional analysis is often preferred by social scientists than longitudinal analysis since the latter includes rather complicated statistical difficulties. Nevertheless, longitudinal analysis is essential because it includes a study of a population over time, whilst cross-sectional analysis is limited to a single point in time. Thus, the most known limitation of cross-sectional analysis is that it does not discriminate between the ageing and the cohort effect. Cross-sectional analysis also involves the limitation of inability of resolving the vagueness in correlations and other measures of associations. Moreover, in cross-sectional analysis the exclusion of essential control variables can lead to false associations. However, one should not always reject cross-sectional analysis since it can be useful when the researcher is interested in the general inclination of some result rather than its dynamic characteristics. Survival analysis [ ... ]