Robert wrote: Copehead wrote:
What makes you say that? You can easily do that and it's being done all the time.
You can statistically analyses a series of numbers not a single number.
What statistical analysis do you think you can apply to a single number, I may have missed that week in lectures and I am always up for a learning experience?
Or took the wrong study altogether. Sure as hell you did not skip that week for reading classes.
Nobody mentioned analysing a single number. The phrasing was ‘perverting’
Of course you can frame a single number by adding vaguely related other figures as a background to which
the said single figure can be made to look good or bad, whatever it is you’re trying to get across.
This is tedious isn't it?
Because he knows that whatever number he gives , you will pervert with statistics.
arguing over whether a plural was used when it obviously wasn't
And we come down to the fear of a whole branch of mathematics because some people don't understand it.
Yomp and stats is like a medieval King and witchcraft
Stats are a tool not of and in themselves bad What Mike is doing is just a way of blanking out anything he doesn't want to here by poisoning the well then shooting the messenger.
It isn't stats that make polling figures wrong anyway, there is little or no statistical analysis to be done, it is sampling, different pollsters sample in different ways and then make fairly arbitrary weighting allowances based on nothing much at all.
That is why a polling company like Survation can get a GE pretty much spot on whilst YouGov were way out even after they realised they were not getting it right.
Because polling figures, as shown by Fiona Bruce's egregious lying on Question time, are used as news items themselves.
But none of this, fascinating as it is, changes the fact that you don't perform statistical analysis on a solitary number as Mike thinks and no amount of dissembling from you is going to change that.