+1
1.0±1.4 kg at a p value of 0.29 is worth discussing?
In the physical sciences test results like these would be laughed out of print, and out of research. I realize medical or life sciences are different, but I can't get my mind away from numbers like - 22+/-52, where the variance is larger, or even double, the mean, or the average result. Such a result is worse than meaningless. Either the test is ridiculous or the researchers are studying the wrong chemicals or processes.
I'm sure you can plug such numbers into statistics calculations but so what can that produce.
"When the search for truth is confused with political advocacy, the pursuit of knowledge is reduced to the quest for power." - Alston Chase
+1
1.0±1.4 kg at a p value of 0.29 is worth discussing?
Few but ripe.
I was interested enough to look into this further. Statistics was never a subject I enjoyed, but in my math studies I was required to take 2 courses of it, and that is the extent of my knowledge of it. But I remember well the basic 'laws' or rules of operations. The result 22 +- 52 means that a 'right' or meaningful test result is within the range of +74 to -30.
Since a result of -30 is 'impossible', that is to say that a test on a live human could not possibly produce such a number, then this number occurs only because of the statistical design and operations performed on the collected data.
Further, I find, that this is commonly accepted practice today, and for some years in the past, to simply delete all such 'impossible' results from the statistical conclusions. (But they continue to believe that conclusions thus drawn are valid.)
NO, NO, NO, NO! Do you hear me screaming at them. Invalid! Not valid EVER.
Worse, this is now commonly accepted practice in all medical tests.
Last edited by Cryptocode; 08-17-2013 at 12:45 PM.
"When the search for truth is confused with political advocacy, the pursuit of knowledge is reduced to the quest for power." - Alston Chase