http://arctic.atmos.uiuc.edu/SEAICE/arctic.historical.seaice.doc.txt
Sea ice extent data is provided by Kelly, et. al. 1988. The ice extent data is compiled for the months April-August for the majority of the period 1901-1956.
The Kelly 1988 data comes from the Walsh 1978 data
http://gcmd.nasa.gov/records/GCMD_WALSH_CHAPMAN_SEAICE.html
http://nsidc.org/data/docs/noaa/g00799_arctic_southern_sea_ice/index.html
1871-1900: A monthly climatology based on Walsh data for 1901-1930 was used, to which spatial variability was added using a 1979-1996 passive microwave bias corrected data set. A climatology based on the passive microwave data defined typical monthly concentrations. Where Walsh data grid cells had concentrations of 100%, and the passive microwave climatology showed concentrations of at least 90%, the climatological concentration was substituted.
1901-1978: Primarily the Walsh data set, with spatial variability added as above. Note that U.S. National Ice Center (NIC) charts weigh heavily in the Walsh data beginning in 1973.
In contrast
Here's some real world measurements from the Russian side of the Arctic taken from published papers explaining those measurements:
Russian historical records of arctic sea-ice extent and thickness extend back to the beginning of the 20th century. There are several distinct periods in the history of Russian sea-ice observations. Occasional ship observations of summer ice edge started in the first decade of the 1900s when the first Russian hydrographic surveys and commercial shipping routes along the Siberian coast began. These data have been analyzed by the Russian climatologist Vize (1944). Some data for this period have also been obtained from Russian navigation books. Starting in 1929, when the Soviet Polar Aircraft Fleet was created, aircraft-based observations began, which improved the quality of the data substantially. However, systematic aircraft and ship observations of sea ice from the Kara Sea through the Chukchi Sea began only in 1932, when the Northern Sea Route was created. There were information gaps during World War II (1942-45). The missing data have been reconstructed using statistical (regression-like) models relating atmospheric processes (SLP gradients and SAT) to ice extent (Kovalev and Nikolaev 1976; Yulin 1990). Aircraft ice-edge observations continued until 1979, when the satellite era began, but until recently a combination of satellite and aircraft summer ice-edge observations was used. Since 1990 all ice-extent observations have been satellite-based.
The choice is between hockey stick ice from a temperature-based model of theoretical ice extent, or non-hockey-stick real world measurements of the actual ice extent.