{"id":1333,"date":"2016-10-25T20:47:03","date_gmt":"2016-10-25T19:47:03","guid":{"rendered":"https:\/\/blogs.bmj.com\/adc\/?p=1333"},"modified":"2017-10-30T13:41:39","modified_gmt":"2017-10-30T12:41:39","slug":"statsminiblog-pants-and-primary-schools","status":"publish","type":"post","link":"https:\/\/blogs.bmj.com\/adc\/2016\/10\/25\/statsminiblog-pants-and-primary-schools\/","title":{"rendered":"StatsMiniBlog: Pants and primary schools."},"content":{"rendered":"

\"20140205-091454.jpg\"I’ve been struggling to get the concept behind random-effects meta-analysis out for some time – it’s the ‘average effectiveness’ in an ‘average population’ – with the prediction interval being the ‘actual width of where the truth might lie’.<\/p>\n

But .. yes .. but what does that actually mean & why does that matter?<\/p>\n

Well.<\/p>\n

Take\u00a0a primary school. Get the average height of the children in each class. Now use that to tell me\u00a0the average height of a child in that school.<\/p>\n

If I make a tonne of jogging pants for them based on the average height, I might\u00a0provide 15% with an OK fit, but mostly they’ll be too long or too short.<\/p>\n

If the data are very very heterogeneous, the average is “true”, it’s just not very useful.<\/p>\n

Does that make any more sense of how awkward a ‘random effects meta-analysis’ result can be in a highly heterogenous meta-analysis?<\/p>\n

 <\/p>\n