This was inspired by some work once done for the Oxford Pain Research Unit. I'd been working out correlations between two independent ways of measuring pain. One such correlation came to 79%, and a colleague commented that this was "about as good as it is likely to get in any biological system (unless you count death and absence of heartbeat, perhaps)". The language used for the analysis is a stats language called R, by the way.
Cartoon Description
Two researchers are looking at a cross-tabulation of hospital patients output by the statistics language R. It tabulates their state HAS HEARTBEAT / HAS NO HEARTBEAT versus DEAD / NOT DEAD. The tabulation shows a fair number of patients who are dead but have a heartbeat, or who are not dead but have no heartbeat. One researcher is saying "I think your tabulation has programming errors"
. The other replies "Naaaah! Biological data never yields perfect correlations!"
.
By the way, the R window reads:
R version 2.19.5 (2013-6-13) -- "Statistician's Heaven"
Copyright (C) 2012 The R Foundation for Statistical Computing
> tabulate.my.data()
CROSS TABULATION OF DEATH VERSUS PRESENCE OF HEARTBEAT IN A
SAMPLE OF HOSPITAL PATIENTS
TOTAL PATIENTS IN TABLE: 96
|
| WITH HEARTBEAT | WITHOUT HEARTBEAT |
-----------|-------------------------|-------------------------|
NOT DEAD | 601 | 368 |
-----------|-------------------------|-------------------------|
DEAD | 268 | 648 |
-----------|-------------------------|-------------------------|