Tuesday, August 26, 2008

Ipsos MORI: Pope - Pointy Hat; Economy - Stupid

Not sure we really need an ipsos MORI poll to inform us that the economy is the major issue of the day. Tom rattles through the soft underbelly of this "PAM" poll.

Immigration - which has now dropped way down - was I think TOP in the same poll in the April edition. The economy was nowhere way back then. Not in the top five as I recall.

Having done quite a lot of field survey work for leading companies my biggest concern is not the figures themselves, or the relatively small samples drawn from a limited number of constituencies, or the accuracy and consistency of the interviewing, or the recruitment of respondents, or yada yada yada. For reasons explained by Bob Worcester linked below these are OK.

Rather I am concerned at polls having a certain amount of predictive power because they can lead rather than follow in terms of both the broad population, some of whom clearly treat voting as a bet on the result rather than a real political choice, and also as agenda setters for commenterati and mass media.

Other long time surveys conducted by ipsos MORI and others - for example on Radio, Television, Newspaper, Periodical numbers - do sometimes show a spike up or down in a single month. These surveys have similar or larger samples than PAM. These rare flash variations can be best explained by sampling idiosyncracies; and they are generally reported only in trade press. That way they don't become self-fulfilling, and they settle down, back to trend.

When it comes to political choices surveys - possibly with the same or greater potential for "noise" - they are reported widely to the broad population and could therefore reinforce sentiment.

Sir Robert Worcester of ipsos MORI explains why the polls are not "all over the place". That is on the basis of almost all of them being clustered within +/- 3% of THEIR OWN AVERAGE. PDF of ipsos MORI tables (66 pp).

No comments: