Readers of this journal often find articles in which an investigator describes the relationship between one dependent variable and one or more independent variables. Some recent examples in the Archives are distance of densest scotoma from fixation (dependent variable) as a function of mean rim-to-disc ratio1; color vision defects (dependent variable) and diabetic retinopathy2; and vitreous penetration ratios (dependent variable) and scotopic electroretinography amplitude.3 In some instances these relationships have become guides to surgical practice: the Sanders-Retzlaff-Kraff regression formula is used to select an intraocular lens that yields emmetropia following cataract surgery (axial length and keratometer readings are independent variables).4 It was formulated from a series of cases of intraocular lens insertions. The purpose of this editorial is to provide a brief guide to the strategy and language of the statistical analysis typically used in these studies.
The investigator assigns a scale of measurement to his
Seigel D. Regression. Arch Ophthalmol. 1987;105(2):185-186. doi:10.1001/archopht.1987.01060020039023