In this study, the comparison of results obtained from the same data set of three alternative probability models known as the Linear Probability Model (LPM), Binary Logit Model (BLM) and Binary Probit Model (BPM) has been discussed. The use of these models is an accurate approach in the event of determining a categorical dependent variable with two levels. In order to throw light on the similarities and differences of the results, data recorded by The Department of Traffic Training and Research of the General Directorate of Security Affairs of Turkey has been examined, and also factors influencing the chance of drivers surviving or not surviving motorcycle accidents during 2002 have been determined.