See>>> http://www.cloudtag.com/
And it’s been ‘validated’!!! >>> http://www.cloudtag.com/2015/10/26/accuracy-results-from-user-testing/
Sorry – correlation is just not the way to validate 2 methods!
They should specify an acceptable margin for the paired differences (2 measures are never identical)
The British Standard is – 95% of all measurements should be with 2SD of the mean difference
Then they should measure the differences they get
That way they quantify the precision of the difference as well by putting confidence limits on it
Mind you no-one ever knows how to do it properly (statistically)!
0 Comments