Aphasia is a language disorder that affects approximately two million Americans. The communication barriers associated with aphasia limit the ability to return to work and hobbies, worsen social relationships, and lead to social isolation. Social isolation is correlated with many negative health outcomes, including mortality. As a result, the consequences of aphasia can be wide reaching and severe. Group treatment has the potential to improve communication and reduce social isolation, while also reducing medical costs. However, this treatment format lacks a strong evidence base. This research uses a hypothesis-driven approach to test the efficacy of conversation treatment and develop and refine a mechanistic pathway of how conversation treatment leads to behavioral changes in communication ability. We will systematically explore whether the effects of treatment differ as a result of number of participants (large group or dyad), group composition (heterogenous versus homogenous profiles of aphasia), and severity of aphasia. The study will enroll a total of 168 participants with aphasia across three sites. In cycle one, 72 participants will be randomly assigned to one of three conditions: dyad treatment, traditional group treatment or delayed control group. In cycles two and three, 48 participants with severe aphasia and 48 participants with mild-moderate aphasia will be randomly assigned to either large group or dyad conditions. In all cycles, treatment will occur for 60 minutes, twice per week for 10 weeks. During treatment sessions, a speech pathologist will facilitate discourse on a focused set of salient topics, such as current events, using individualized, linguistic and multimodal cueing hierarchies. The primary outcome measure is a functional measure of communication (Aphasia Communication Outcome Measure, ACOM; Hula et al., 2015). Secondary outcome measures will include standardized tests and patient reported outcome measures, and examine linguistic and functional communication abilities. All participants will be assessed pre-, post-, 4 weeks post- and 12 weeks post-treatment. The proposed research will address the following specific aims. Specific Aim 1 is to examine efficacy of conversation treatment delivered in dyads and large groups compared to a delayed-treatment control group. Specific Aim 2 is to determine optimal parameters for conversation treatment by testing whether the treatment effects differ as a function of aphasia severity (2A) or group composition (2B). Specific Aim 3 will test a hypothesized model of the pathway by which conversation treatment effects behavioral change. The results will provide further efficacy for conversation treatment and inform about the optimal parameters and outcomes of this intervention. This work takes a vital first step towards elucidating the mechanisms of change in conversation group treatment, with a long-term goal of ensuring access to cost-effective care for people with aphasia.