http://www.sepp.org/NewSEPP/Kyoto-Patterson.htm
""
However, the scientific review process that all studies must undergo before publication had failed in the case of the MBH98 paper. The temperature data before 1900 were not directly measured, as they were after 1900 when land-based thermometer readings were used. Instead, pre-1900 temperatures were calculated based on the measurement of "proxies," natural phenomena such as the growth of tree rings or coral that indicate what temperature was at certain times in the past.
Consequently, grafting the two very different types of data sets together without significant overlap to come to dramatic conclusions was unwarranted and should have been seriously contested by the paper's reviewers.
Chris de Freitas of the School of Geography and Environmental Science at the University of Auckland, New Zealand, sums up, "The Mann 'hockey stick' is nothing more than a mathematical construct vigorously promoted in the IPCC's 2001 report to affirm the notion that temperature changes of the 20th century were unprecedented."