50% of all...

To be clear from the outset - this is not a post about "50% of all trials are unpublished" so calm, everyone... 50% is a very common percentage, apparently. I was reading a Guardian article this week and was increasingly distracted by the frequency with which "50% of all..." was occurring.

The article theme was that pharmaceutical industry/NHS collaborations were not necessarily all bad, and there were many examples of where specific initiatives were benefitting local services. However, the frequency of "50%" statistics was frightening:

• Participants at the roundtable agreed that action was needed to tackle the fact that 50% of patients on medication do not take it properly
  
• Even some patients on life-saving cancer medication do not take it, stressed Barber. "There's an ignorance about this. About half [of the non-taking] is intentional and half is non-intentional
  
• ...working with the South Essex Partnership University NHS foundation trust to improve the management and care of people with mental health problems, such as by spotting early signs of a relapse of schizophrenia. It has led to a 50% drop in hospitalisation rates among such patients, she pointed out.
  

So perhaps it is not so surprising that "50% of all..." is often confused by people picking either no evidence (as was the case in the Guardian article) or the wrong evidence (sloppy journalism), or the interpretation of these statistics is blurred by fudging the general theme of an article together into a single zombie?
And no, there were no citations for any of the 50% statistics in the Guardian article.