On 31/03/2012 12:12, Roland Perry wrote:
In , at 11:42:22 on Sat, 31 Mar
2012, Graham remarked:

Do such figures have any credibility whatsoever?

Have you ever been involved in market research? They don't just pull
numbers out of a hat.

They might as well for some of the results they get. This seems
particularly likely if the "research" has been commissioned just
to provide an excuse for a press release and hence some free
publicity or if the questions given to the MR firm dictate what
answers are acceptable (as with the drinking survey Arthur saw not
permitting a choice of real ale).

This is a typical bit of Usenet nonsense, where people are too inclined
to rubbish the efforts of other professionals, while claiming their own
activities are a tour de force that's beyond reproach from mere
amateurs.

As in the Titanic was built by professionals, the Ark by amateurs?

I wonder how many marketing/PR people have more than a GCSE in maths?

Even aside from that, I have been involved in checking over some market
research questions compiled by a dedicated department in a big company,
and they were useless for finding data. Some things are just amusing:
"location .... f) rest of world g) other" is a common one. Some make
interpretation difficult: asking people to rate things without saying
whether 1 or 10 is best, or even swapping part-way. We found our
customers gave our widgets a good rating for all except one factor,
where they got a bad rating from almost everyone. This was the only
question where 1 was the option to tick if you were happy, rather than 5...

Nor would they claim that a figure like "22%" is
accurate between 21.9% and 22.1%, but 22% is an awful lot of people
(about 15 million) and needs to be taken seriously.

The 22% figure relates to over-65s.

Yes, a silly mistake. I should have used the "12% figure". (That's still
more than the entire population of London).

Another characteristic is the odd non-sequitur that gets thrown
in. 40% of people taking time off during the games, eh? Wouldn't
have anything to do with the games occurring during the main
holiday season, would it?

What is the sample - people who have booked holidays recently and so
appear on holiday firms' databases of people to be contacted?

I did wonder about that myself, but would 40% of the population take a
holiday during a normal mid-August? Seems very high, looking around the
places I've worked in the past (and we know people without schoolkids
try to avoid August because of the prices).

But might find they have to go on holiday then, as it is a quiet time so
they can be spared (eg all their French customers have vanished). Or
their holiday year runs to August, so they have to use up some odd days.


I'm quite happy to accept that the 40% *includes* the 12% and the 9%.

And that when they say "time off" they mean "annual vacation", not rest-
days, weekends and so on.

The EU have done some statistics which say that (taking the population
aged over 15) about 57% of Brits take a holiday in any one year (2008),
with 29% exclusively going abroad, 20% exclusively staying in the UK and
8.7% doing a bit of both.

http://epp.eurostat.ec.europa.eu/cac...-10-010/EN/KS-
RA-10-010-EN.PDF

Of course "taking time off", and "going on a holiday" aren't the same
thing.


--
Arthur Figgis Surrey, UK