This study compared cortisol concentrations yielded using three saliva collection methods (passive drool, salivette, and sorbette) in both in vitro and in vivo conditions, as well as method acceptability for a sample of children (n = 39) with High Functioning Autism Spectrum Disorders. No cortisol concentration differences were observed between passive and sorbette samples obtained in vitro or in vivo. The salivette derived concentration was lower than the other two methods for the in vitro derived comparisons but did not differ from the other methods when collected in vivo. Cross-day comparison for the salivettes was also found to differ significantly, whereas the cross-day comparisons did not differ for the passive method or the sorbette method. Overall, passive drool and sorbettes were found to produce similar and stable readings of cortisol, whereas the salivette yielded unstable and variable concentrations. Ratings suggested that the children generally perceived all methods as acceptable.