This study evaluated the efficacy of a computerized behavioral skills training (CBST) program for teaching responses to interview skills by adolescents and young adults with learning disabilities. The package consisted of instructional videos, video modeling, rehearsal, feedback, and selection-based instruction. A multielement design was incorporated within each tier of the multiple-probe design such that each question was pseudorandomly assigned to a CBST condition or a CBST plus selection-based instruction condition. Data suggest that the CBST program was effective in promoting accurate topography-based responses to interview questions regardless of whether selection-based instruction was included. However, CBST plus selection-based instruction may have been the optimal condition for 4 out of 5 participants.