Why not subscribe?

Saturday, October 20, 2007

Do you hate Brits? You will after you read this.

Binyamin Dissen wrote:
> http://news.yahoo.com/s/nm/20071019/od_nm/britain_popularity_dc_1
>
> "A poll of Europeans showed people of different nationalities liked
> each other more after getting to know each other, except in the case
> of the British -- who became less popular."


* WOW! This could be a whole chapter in "How to Lie with Statistics". Let's
look at this.

"BRUSSELS (Reuters) - A poll of Europeans showed people of different
nationalities liked each other more after getting to know each other, except
in the case of the British -- who became less popular."

"A project organized by the Notre Europe think tank brought together 362
citizens from 27 EU states for two days of deliberations in Brussels last
weekend."


* 362 citizens, picked how? Even if they were picked randomly, it's unlikely that they agreed to participate randomly.
* 362 citizens, 27 states -- so there are maybe 12 Brits.
* There's probably no country in earth that doesn't contain 12 people I'd
dislike if I met them.
* Brits from where? England, Wales, Scotland, Northern Island? Are Brits that uniform?
* What were they deliberating on? Did it have anything to do with issues on
which the British position may be different than the continental position?
(currency, trade, European constitution, etc.) As it turns out, it
probably did. On Professor Fishkin's web site, he describes the method as
follows:

http://cdd.stanford.edu/polls/docs/summary/

"A random, representative sample is first polled on the targeted issues.
After this baseline poll, members of the sample are invited to gather at a
single place for a weekend in order to discuss the issues. Carefully
balanced briefing materials are sent to the participants and are also made
publicly available. The participants engage in dialogue with competing
experts and political leaders based on questions they develop in small group
discussions with trained moderators. Parts of the weekend events are
broadcast on television, either live or in taped and edited form. After the
deliberations, the sample is again asked the original questions. The
resulting changes in opinion represent the conclusions the public would
reach, if people had opportunity to become more informed and more engaged by
the issues."

* So, basically, this is a sort of focus group. It's an interesting
technique, but it is easy to see that some of the Brits might have had
different viewpoints (at the beginning or the end) that might be at
political variance with the rest of the EU.

"They were asked their views on a range of issues before and after the
event, including how much they liked or disliked German, French, Polish,
Italian, British and Spanish people."


* Why these? Is Poland the 6th largest country of the 27?

"The Spaniards were most popular with 78.6 percent approval at the end of
the weekend. The Poles the least with 67 percent."


"But all nationalities gained in popularity, bar the British who went from a
70.3 percent approval rating at the start of the weekend to 68.1 percent by
the end."


* Assuming independence, this difference is not even close to being
statistically reliable. The difference is .022 and 2 standard errors is
.069. It's possible that with repeated measures this MIGHT be significant,
but it seems unlikely. I can't find more detail on Fishkill's site about this particular study.

"Professor James Fishkin of Stanford University in the United States, who
developed the polling technique used, urged some caution about the
findings."


"I would be careful about drawing too much inference about people disliking
the British -- it's small and not significant, but you know, it is what it
is, and it did happen that way."


* Oh, wait. Buried deep down here is the fact that it is small and
nonsignificant. So who do we blame for the fact that this got such press?
The PR people for Fishkin or the conference? Reuters? Our own gullibility?
I don't know.

"Fishkin said he did not think France's defeat by England in the rugby World
Cup semi-final in between the two polls had had an influence, even though
the number of French participants in the survey was disproportionately
high."


* Ah, a plausible alternative explanation.

* So, after listening to my discussion, do you hate Brits more? Probably you don't, but if Reuters can use a misleading headline, why not me?