We applied multiple stepwise logistic regression analysis to the preoperative data for 153 glaucomatous patients to determine the predictive value of different variables in the outcome of laser trabeculoplasty. One hundred patients had capsular glaucoma and 53 patients had simple glaucoma. Of the 14 preoperative variables included in the analysis, the following best predicted successful outcome of the laser treatment: laser trabeculoplasty performed as primary therapy, low maximum pretreatment intraocular pressure, a diagnosis of capsular glaucoma, no preoperative treatment with pilocarpine, and the operator's experience with the procedure. Age, sex, pigmentation of the angle, the energy level used, the number of burns, the use of preoperative medication other than pilocarpine, and the occurrence of an immediate postoperative increase in intraocular pressure could not be shown to have predictive value as to the outcome of the treatment. Inclusion of postoperative follow-up period did not improve the model.