Introduction

A special type of regression known as Tobit regression is used to handle the truncation at zero. This regression is used for the analysis of social support income, defined as the sum of Canada Pension and 'income from other government sources'. Tobit regression assumes the data are truncated, or censored, above or below certain values. The technique is similar to survival analysis in that it uses a censoring value to stand in for an indeterminate quantity. In fact, it is run using one of the procedures in SAS designed for the analysis of survival data.

The parameter estimates resulting from a Tobit regression are not interpretable in the same way that those from a linear regression are. Techniques are available which decompose the parameter estimates in two parts: the effect of the probability being above zero, and the effect on the mean income given that it is above zero.