Roger Coppock wrote:
CO2 or Sunspots: Statistical Correlation Chooses

Statistical correlation is a powerful technique with
very many uses. It produces "R squared" a measure of
whether two series of measures trend together.

(Those who are new to statistical correlation and
"R squared" will find a tutorial on the subject he

http://mathworld.wolfram.com/Correlation.html

http://mathworld.wolfram.com/Correla...efficient.html

Item 20 in the above shows R squared for several graphed
relationships.)

When applied to a time series of global mean surface
temperatures and data from prospective global warming
causes covering the same time period, correlation can
help locate the cause of the observed global warming.
Low "R squared" values, those near zero, can, by
themselves, totally rule out a prospective cause.
High "R squared" values indicate that a prospective
cause is very likely, but do not, by themselves,
'prove' something caused the warming. (Experimental
science rarely 'proves' something like a mathematical
proof does.)


Below are directly observed data for global mean surface
temperature, CO2 concentration, and sunspots for the last
50 years. This is as long as the longest directly
observed record of atmospheric CO2 concentration.
...

As originally posted in spawning thread:

Why only 50 years? I'm not considering CO2 as part of my correlation,
why are you?

This is an arbitrary and fictitious limit that you have imposed without
sound justification, and therefore I must conclude that your results are
(deliberately?) skewed.

If you wish to engage in the discussion with your 'winning
computations', please use a timespan that is equivalent to what we are
discussing, specifically 1850-2000.