The aim of this work was to test the comparative efficacy of four therapeutic modalities (self-exposure, self-exposure + Alprazolam, Alprazolam, and self-exposure + placebo) and also to determine the combined effects of self-exposure with Alprazolam and self-exposure with placebo in the treatment of agoraphobia without current panic. The sample consisted of 31 patients selected according to DSM-III-R criteria. A multigroup experimental design with repeated measures of assessment (pre-treatment, post-treatment and 1, 3 and 6-month follow-up) was used. The results indicated that there was a similar therapeutic improvement (in about 75% of the cases) between pre- and post-treatment in all therapeutic modalities, except for the Alprazolam group, where improvement did not take place, was rather weak or tended to fade as time passed. This improvement increased at the follow-ups in the self-exposure + placebo group, remained stable in the self-exposure group, and was irregular or fairly unpredictable in the self-exposure + Alprazolam group. There was a positive combined action between self-exposure and placebo and a negative interaction between self-exposure and Alprazolam. The highest relapse rate appeared in the therapeutic modalities where the active drug was administered. The intratreatment evolution was faster in the self-exposure group than in the others, but it tended to remain stable in the second part of the therapy. It is therefore concluded that the efficacy of self-exposure therapy may be the same if reduced to half the number of sessions. Finally, several topics that may contribute to future research in this field are commented upon.