Canadian researcher John Coates took saliva samples in
the morning, twice a day for eight days, from 17 men working on a London
mid-size trading f loor (trading a wide range of assets, with largest exposure to
German interest rate futures), in June 2005, and classified each trader according
to whether his testosterone level was high or low on that day (compared
with the trader’s median over the period). High testosterone days differed from
trader to trader, and high days differed from low days on average by 25% in
testosterone level. He also recorded the profits or losses (P&L) in pounds sterling
of each trader during 11 am–4 pm daily.