Stats and Polls Letter

Scotland on Sunday published this letter of mine today. Thankfully we have been spared these polls for the last two weekends, but no doubt this process of dubious manipulation will be in evidence again shortly.

In all these polls, as with letters written to a paper, the key is always to scan down to check who has commissioned it or written it, before actually reading the text. That way, you get a more realistic perspective and know where the story is at least comingfrom first.

DOUBTS OVER VIOLENT POLL SWINGS
Many of your readers will be wondering, and, indeed be suspicious of why political opinion polls can change so drastically within a few weeks and then seemingly change back again for little apparent reason. Ten point leads cannot disappear and then be re-instated without some-one, somewhere manipulating the methodology or the exact questions asked. Voters are not so engaged in the political process that individual policy announcements can have this sort of effect. It is both a statistical and political impossibility.
There is no consistency in the methodology of these polls. In the last few weeks, Undecided voters were included or not, presumably on the basis of what would give the better outcome. It appears that the party commissioning the report can suggest some of the parameters in the analysis, and it appears that neither print nor broadcast journalists are sufficiently wise to this. Either that, or the roller coaster is much more interesting to them, and this charade is allowed to continue for this reason.