Background: Clinicians often modify evidence-based psychotherapies (EBPs) when delivering them in routine care settings. There has been little study of factors associated with or implications of modifications to EBP protocols. This paper differentiates between fidelity-consistent and fidelity-inconsistent modifications and it examines the potential influence of two clinician characteristics, training outcomes, and attitudes toward EBPs on fidelity-consistent and fidelity-inconsistent modifications of cognitive behavioral therapy in a sample of clinicians who had been trained to deliver these treatments for children or adults. Methods: Survey and coded interview data collected 2 years after completion of training programs in cognitive behavioral therapy were used to examine associations between successful or unsuccessful completion of training, clinician attitudes, and modifications. Modifications endorsed by clinicians were categorized as fidelity-consistent or fidelity-inconsistent and entered as outcomes into separate regression models, with training success and attitudes entered as independent variables. Results: Successful completion of a training program was associated with subsequent fidelity-inconsistent modifications but not fidelity-consistent modifications. Therapists who reported greater openness to using EBPs prior to training reported more fidelity-consistent modifications at follow-up, and those who reported greater willingness to adopt EBPs if they found them appealing were more likely to make fidelity-inconsistent modifications. Conclusions: Implications of these findings for training, implementation, EBP sustainment, and future studies are discussed. Research on contextual and protocol-related factors that may impact decisions to modify EBPs will be an important future direction of study to complement to this research.