Let me explain by quoting Sokal and Rohlf the difference between regression
and correlation. It is funny that you did not know the similarities. Some
researchers say the two methods are synonymous, but it is how the two
methods are used that makes them different.
"...earlier texts did not make the distinction between the two seperate
approaches sufficiently clear....At least one current text synonymizes the
two, a step that we feel can only compound the confusion....[It] is the data
available for analysis may be such as to make one or the other technique
inappropriate." [Biometry, 2nd ed., Freeman, 1981]
The regression method is used where there are independent and dependent
variables. Month is not dependent on temperature and temperature is not
dependent on month. Month does not cause temperatures. Whereas irradiance
and GHGs cause temperatures. Therefore the data that I correlated, although
they covary, there are not caused by each other. Regression is used to test
hypothesis, correlation is used to determine if there is an association
between variables that may or may not be dependent. If month causes
temperatures, that would be even funnier than your sudden revelation that
the two methods have the same R
chao,,
>By the way, I did a linear regression with temps as the dependent variable
>and time as the explanatory variable for January.
>
>The mulitple R: 0.23716
>Pearson's Product Method:0.23716
>
>Well I be damned. Same damn number. Well what do you think this means.
>How about that John is essentially doing a linear regression. The Pearson
>Product Moment Correlation measures the linear relationship between two
>data sets.
>
|