Extended contact has been shown to improve explicit and implicit attitudes toward a number of outgroups, but not yet toward people with mental health conditions. Using people with schizophrenia as the target group, this experiment is the first to demonstrate that extended contact can reduce explicit prejudice, buffer stress responses to future interactions, improve non-verbal behavior, and improve the quality of interactions in a manner detectable by the target group member. Participants watched a video of a brief, positive interaction between two strangers, one of whom they were led to believe had schizophrenia. Control participants watched the same video without being told that the person had schizophrenia. They then participated in a social interaction with a confederate whom they were led to believe had the disorder. Participants' cardiovascular and electrodermal activity were monitored immediately before the interaction. The interaction was also secretly recorded to allow independent judges to assess the participants' non-verbal behaviors. The confederate also rated the positivity of each interaction. Participants in the extended contact condition reported more positive attitudes toward people with schizophrenia, displayed more positive non-verbal behaviors, and had a more positive interaction with the confederate. Moreover, just prior to the interaction, participants in the extended contact condition displayed smaller anticipatory stress responses, as reflected in smaller changes in interbeat interval and non-specific skin conductance responses during this phase. Together, these findings support the use of the extended contact as an intervention that could lead to genuine changes in attitudes toward and treatment of people with severe mental health disorders.