07-02-2009, 08:01 PM
Has anyone looked at why this "eleven year" figure seems to come up a lot? It seems like a very strange interval to use. It's not a moving average, it's not a round number, it's just an arbitrary period. So why pick that one?
It's the oldest trick in the "lying with statistics" book: pick your interval carefully. Measuring from peaks to troughs shows declines, and measuring troughs to peaks shows increases. 1998 was an anomalously warm year, so any trend measured from that peak is going to appear to go downwards. It depends heavily on which data set you use, but however you slice it, always starting in 1998 is going to perpetually skew results downwards.
-Jester
It's the oldest trick in the "lying with statistics" book: pick your interval carefully. Measuring from peaks to troughs shows declines, and measuring troughs to peaks shows increases. 1998 was an anomalously warm year, so any trend measured from that peak is going to appear to go downwards. It depends heavily on which data set you use, but however you slice it, always starting in 1998 is going to perpetually skew results downwards.
-Jester