Previous research has demonstrated the utility of using lag schedules of reinforcement to increase response variability of children with autism. However, little research has evaluated whether the lag schedule promotes variability from within an already‐established repertoire or expands the current repertoire by promoting the use of new responses (i.e., those not previously demonstrated). Thus, the purpose of the current study was to evaluate the extent to which lag schedules of reinforcement produced already‐established intraverbal responses or novel responses for 3 children with autism. Results showed that lag schedules alone were sufficient to increase the number of different responses emitted for 2 participants, whereas brief variability training was needed for 1 participant. Further, some participants emitted novel responses throughout the experiment, suggesting that lag schedules may be an effective method for expanding a response class.