MODELING OF DISCONTINUOUS RELATIONSHIPS IN BIOLOGY WITH CENSORED REGRESSION

Discontinuous relationships between variables are common in biological data. Discontinuities can sometimes give the appearance of curvilinea rity, which suggests the data should be analyzed with nonlinear models . Here we show that often a more meaningful analysis can be obtained w ith censored regression techniques. In a censoring model all points be low (or above) a certain threshold are observed only by the value of t he threshold (e.g., a baseline temperature). We illustrate the method with an example from plant reproductive biology: plant reproductive ma ss is never negative but becomes positive only after some ''capacity'' to flower reaches a threshold. The vegetative mass at which the thres hold is reached and the relationship between reproductive mass and veg etative mass above the threshold are estimated from data. Using censor ed regression with real and simulated data shows that apparent curvili nearity suggested by models that do not account for censoring can be a n artifact.