Stimulus equivalence is a teaching paradigm with empirical evidence for the establishment of a variety of skills (i.e., letter and number recognition, sight word reading, face-name recognition, etc.) in typically developing and non-typically developing children and adults with different levels of functioning. Simple and complex conditional discrimination training have both been demonstrated to be effective. However, the effectiveness of the two procedures has not been directly compared. The present study investigated the relative effectiveness of these two procedures to establish sight word reading and rudimentary reading comprehension to one typically developing children and two children with autism. An adapted alternating treatments design was implemented, whereby stimulus sets were assigned to either a simple-sample or complex-sample condition. The percentage of correct responses was scored for each training condition, and the number of trial blocks required to reach criterion was compared to assess the efficiency and effectiveness of the two conditions. Generalization probes were conducted for all three participants, and maintenance probes were conducted two weeks following the end of training for two of the three participants. Results indicate that complex sample training was more efficient than simple sample training with all three participants. Also, all participants scored higher on their post-training probes compared to their pre-training probes in both conditions which demonstrates utility of stimulus equivalence.