Monday, November 21, 2011

Hoy! Poll Oy!

Apple Daily poll above. Weirdly, Ma seems to have increased his percentage and suffered only a tiny drop in his lead despite the steady drumbeat of bad news that is killing him in other polls -- the prediction market has Tsai over Ma 48-44 when asked who will win, as of this writing. This Apple Daily poll appears to be badly wrong. The poll also shows Ma with double digit leads in the Taipei basin, Taoyuan/Hsinchu/Miaoli, and central Taiwan. Say what? Because the Taipei Times reported today that the internal polls showed Tsai ahead in central Taiwan.
Recent surveys showed that the support rate of DPP Chairperson Tsai Ing-wen (蔡英文) in central Taiwan — Greater Taichung, Changhua County and Nantou County, which is sandwiched between the Jhuoshuei and Da-an rivers — has surpassed that of KMT candidate President Ma Ying-jeou (馬英九), a senior aide of Tsai’s campaign said.

The Chinese-language magazine the Journalist on Wednesday quoted an anonymous KMT official as saying that an undisclosed KMT survey showed that Tsai is leading Ma in central Taiwan by 10 percentage points, and by 2 to 3 percentage points overall. The aide also said that Tsai had overtaken Ma in central Taiwan, a region with about 2.4 million voters.
The NCCU prediction market, broken out by area, also has Tsai up slightly on Ma in the central area  [MT error: that is only for the Hakka vote]. UDN also put forward a poll of younger voters (20-29) which had Tsai and Ma tied at 39 apiece.
Don't miss the comments below! And check out my blog and its sidebars for events, links to previous posts and picture posts, and scores of links to other Taiwan blogs and forums! Delenda est, baby.


Anonymous said...

A couple of thoughts. Apple is not my favorite poll for this simple reason that they use automated interviews rather than actual humans. I don't know how this affects results, but I find it hard to believe it doesn't introduce some sort of selection bias.

I'm also not a big fan of the NCCU futures market for the simple reason that they don't use real money (because that would be illegal gambling). The Iowa futures market, which is well respected, does use real money. In the NCCU market, you can vote with your heart and pay no real price. If there were real money at stake at some of the prices quoted, I'd certainly be in there cleaning house. But why bother?

I wouldn't be that surprised if Ma did get a small bump last week. It wasn't all bad news for him. The party list got a lot of positive response. However, I expect that bump to fade as that story recedes and newer ones (like the Robin Hood narrative and the gambling capo meeting) take over.

paul said...

just goes to show you cant really trust polls, especially the ones in taiwan.

Anonymous said...

Michael, that's the *Hakka* vote broken out by region. It's also ridiculously low volume and the total sum of the values > 100, which means the market doesn't even agree with itself on a consensus. Not that it's probably any worse than the polls out there on this...

By the way, have you considered the new Blogger layouts? Some of them are a really big jump forward in usability and aesthetics in my humble opinion.

Michael Turton said...

I find it hard to believe it doesn't introduce some sort of selection bias.

Thanks, this isn't the first weird Apple Poll. Now I understand better.

Michael, that's the *Hakka* vote broken out by region. I

Ha thanks. I'll fix that.

Anonymous said...

I meant the NCCU market link was a Hakka vote breakdown, not the poll at the top.

RollingWave said...

It would be pretty hard to frame Apple as pro-blue, in domestic politics here it's fairly neutural if you look at their editorials which randomly features views and articles from both sides, but the founder, despite of (or maybe because of ) being a Chinese who fled from Guangdong in the worst days of the Mao era and snuck into Hong Kong, has been quite well known of being outspokenly anti-CCP since the Tain An Men protests.

The future's exchange had Tsai and Su winning in last year's mayor races.. so it's not exactly the holy grail of predictions, while the AIT polls are unconfirmed and there's no detail on their previous predictions (how many polls they did? was all their polls pointing to the same result all the way? etc..)

In short, unfortunately the underground gambling rings might actually be the most accurate predictor . since as others have pointed out, people don't joke with their money no matter their leanings. espeically not gangsters that run the rings.