In the study, we compared the effectiveness and efficiency of two error correction (EC) procedures --a model of the correct response (MoCR) and remove and re-present (RRp)--when using discrete trial training (DTT) in teaching to identify the community signs to children with autism spectrum disorder. We also examined the maintenance and generalization effects on children's behaviours and the preferences of neurotypical peers about the EC procedures. Four children participated in the study. We used an adapted alternating treatments design. The findings demonstrated that MoCR and RRp did not differ significantly in terms of their effectiveness, although DTT with MoCR was more efficient than DTT with RRp for all participants. Moreover, the findings revealed that all participants had maintained the behaviours, while three of the four participants were able to generalize. The neurotypical peers had a more definite preference for MoCR over RRp. Implications for future research were discussed.