3 ms·
I'm always suspicious of "percent change" data that isn't supplemented with absolute values. If there was a big change for the worse, does that mean they used t
by extension 15y ago
I'm always suspicious of "percent change" data that isn't supplemented with absolute values. If there was a big change for the worse, does that mean they used to be good and are now normal or that they used to be normal and are now bad? And is the value high enough to even make the change meaningful? You have the same problem with any other change. It just seems to raise more questions than it answers.
- wycats 15y agoYou can get some raw data from Wolfram Alpha: http://www.wolframalpha.com/input/?i=United+States%2C+United+Kingdom%2C+Canada%2C+Russia%2C+France+|+diabetes http://www.wolframalpha.com/input/?i=United+States%2C+United...