A quick word on the last polls

The polling blackout starts tomorrow, so legally no one can publish or publicize poll results. It’s unclear whether that affects bloggers like me. If I were based in the USA, I’d gleefully ignore the law. However, I live and work in Taiwan, so maybe I should obey the law. I consider it more of a silly inconvenience rather than a violation of my fundamental rights, so it isn’t exactly bowing to tyranny to stop talking publicly about polls for a few days.

Anyway, several polls have been published in the last few days, and these polls seem to go in different directions. Some show that Tsai has dropped five points or so, others show Soong picking up significant support and being almost even with Chu, still others show very little fundamental change from the last two months, and the KMT poll is all alone in showing Chu trailing Tsai by less than 10%. In general, I am somewhat trusting of media polls and pretty skeptical of polls released by candidates  or parties. This year, especially as we have gotten later in the campaign, more and more polls have set off my bullshit alarm. Lots of them have seemed to be very politically useful for a specific candidate. We’re getting these sorts of polls in legislative races too, though polls for individual legislative districts and party list votes are intrinsically more volatile just because the question isn’t always as clear. (One poll showed the NPP getting nearly 10% of district legislative votes. Remember, they only have four credible district candidates. Even if those four all won overwhelming victories, the NPP wouldn’t get anywhere near 10% of the national vote. I think some respondents were thinking of the party list vote.)

What’s my interpretation? I think the debates probably gave a tiny bump to Soong and a tiny nudge downward to Chu and Tsai. However, I suspect any effect from the debates will recede as the debates fade into memory. Over the last 11 days, I suspect public opinion will revert to the longer-term equilibrium unless something new happens to upset it. So I still see this as roughly a 45-30 green-blue split in polls, which translates into roughly a 57-42 split in votes. However, the division of the blue votes between Chu and Soong is still a little unstable. (You may have noticed that Chu has spent almost as much time and energy trying to shore up his deep blue support as going after Tsai.)  In addition, turnout will probably be higher on the optimistic green side than on the relatively demoralized blue side. This is roughly how I’ve seen the election developing for at least a month now. In short, I’m not paying too much attention to the final polls, other than to make sure I don’t see anything credible that signals that something fundamental has changed. Thus far, I haven’t seen that type of thing.

6 Responses to “A quick word on the last polls”

  1. ジェームス (@jmstwn) Says:

    The lack of legislative district polling has been a huge disappointment. Do you think TVBS hasn’t polled districts because of budgetary constraints? because of lack of interest? or because the results it’s getting would cause panic in the blue camp?

    • frozengarlic Says:

      TVBS didn’t do a lot of polling at all this year. I’m not that surprised they didn’t do much legislative polling. There were some district polls that slipped out, but not too many. I just don’t think the public was all that interested in individual races. Most people just want an overall number.

  2. Julian Says:

    The LT poll showed less than 5% support for Chu among 20-29 y.o. that’s pretty shocking, no?

    • frozengarlic Says:

      Remember, in a normal sample of 1000 respondents, the sampling error is about 3%. However, of those 1000 respondents, fewer than 200 will be in the 20-29 demographic. That means it has a sampling error of 7% (or even larger if it is significantly less than 200 respondents). I’d focus on the overall level of Chu’s support in the full sample, which is shockingly low in and of itself. It’s probably lower among young voters, but we really can’t tell how much lower with that data.

  3. buckhead Says:

    Thought might as well share here. I recalled Mr.Garlic wrote about the same subject before. What disappoints me is they don’t have upcoming events integrated in here yet.

    https://relab.cc/candimap2016/#/

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: