Okay, I have nothing like the expertise of some of you. It is very impressive and you really know the subject.
Statistical correlations and significance are expressed as a number between 0 and 1 with zero being no confidence and 1 being 100% correlation. (actually this is between-1 and +1 but we are not dealing with negative correlations are we?)
I cannot find these numbers for the usual subjects in this category.
1.What is the correlation number for rise in temperature compared to CO2 in the atmosphere for a period greater than the last 100 years? (it is acceptable to break this down into 50 year increments because of nonlinearity)
2. What is the correlation number for fossil fuels burned versus increase in atmospheric CO2 over the last 150-200 years, or even the past 60 if the data are not available?
No one ever cites these tests for significance in their arguments. This question was stimulated by an answer to question saying the numbers had been put into an SPC program and showed temperature to be "in control" Applying industrial statistics is not going to show a lot in this area because those are generally tests for stability in a system and "in control" implies some sort of chart like IMR was used.
I don’t care if the method used was linear regression (for univariate) chi squared, ANOVA (for bivariate?) whatever. But if you know the tests please include them if not don’t worry about it.
These are numbers we should all all be able to recite off the tops of our heads. E.g. the correlation between temperature rise and increased atmospheric CO2 is XX%
Simple bivariate analysis.Levels of confidence. Not complex explanations.
Whoa! don’t calculate this yourself. This should be in the peer reviewed research. The academics should have done this already as a part of their analysis. Including the compensations for the fact that the system is non linear.
BTW: I agree the .78 for 48 years is surprising, given the variables and non linearity.
Robert-I agree those numbers are extraordinarily high I also agree linear regression is not a proper test.
Thanks for the radiocarbon dating explanation, I already knew this but you did a good job of explaining.
I guess I am surprised no one can just cite a particular piece of research. It would seem to be cut and dry. From the physics of the greenhouse effect, to human CO2 emissions,, to increasing temperatures, someone should have done the analysis to show that 1) This is not a part of some "natural" cycle because the rapidity of change is so great. And 2) it is directly related to human activity.
Someone had to have done this analysis???
Technorati Tags: 100 years, academics, anova, atmospheric co2, bivariate analysis, btw, compensations, correlation, explanations, fossil fuels, increments, industrial statistics, linear regression, negative correlations, nonlinearity, quot, spc program, statistical correlations, temperature rise, variables