We originally used 58 subjects over 4 sessions. After looking at the data using rudimentary techniques (similiar to those used in section 5.2), we noticed clearly 1) estimated risk parameters were equal to estimated rho parameters. There was some correlation between risk parameters being less than eis parameters, and a choice of early resolution. We attempted to double our subject size to see if this effect would persist. It did. We ended with 101 subjects.The were no formal power cacluations done as we were largely contrained by financial resources.