From: Mike Jr on

http://noconsensus.wordpress.com/2010/01/19/long-record-ghcn-analysis/#more-7578

"Raw Data

So what story does the raw data from the 1034 long record GHCN
stations tell? Shown below is the unweighted mean of the temperature
anomalies for all data from those stations for each month from 1999 to
2005. The mean is slightly smoothed, hence the high Lag-1 serial
correlation. The trend is fairly low, at 0.0269 Deg. C /Decade – which
are the units I hereafter use for all trends. The confidence interval
is over three times as high as the trend, which is accordingly far
from being statistically significant. There is no pronounced peak in
1998, and the peak temperature occurred in the 1930s. As this data set
largely represents temperatures in the USA, that is perhaps
unsurprising."

Adjusted data
"In order to get a clearer picture of the effects of the adjustment
process, I calculated the trend in the combined raw data from all 764
stations which had long records of adjusted data, thereby ensuring a
like-for-like comparison. The mean raw temperature anomaly record of
those 764 stations is shown below.

What this shows is that on average the adjustment process more than
quadrupled the trend in raw temperatures, increasing the trend of the
mean from 0.0113 for raw data to 0.0536. Indeed, if the mean trend
change of 0.0423 resulting from adjustments to the long record
stations were typical of the effect of adjustments to station data
generally, the adjustment process would account for a substantial
proportion of the recorded global mean temperature increase over the
twentieth century."

--Mike Jr.