One method of conducting preference assessments with individuals who have autism involves measuring the time spent by the child engaging with various stimuli. Engagement is generally defined as showing interest in the stimulus but few studies have investigated the potential effects of the quality of engagement with the stimulus on reinforcer effectiveness. In this study, both the duration and quality of engagement were measured for four children with autism aged between 4 and 5 years who participated in a duration-based multiple stimulus with replacement preference assessment. Stimuli were classified into one of the following three categories based on duration and quality of engagement with each stimulus presented: no interest; interest and; active engagement. The relative reinforcing value of stimuli from each category was then investigated. Results showed that the quality of engagement with the stimulus did not predict the relative reinforcer effectiveness of the selected stimuli for any of the participants. For two children, the procedure was effective in differentiating stimuli classified as no interest and those with which the child had engaged, regardless of the quality of that engagement. The utility of an engagement-based preference assessment that also measures quality of engagement with stimuli as a means to guide selection of reinforcers for children with autism is discussed. (Contains 2 figures and 4 tables.)