Voter Education: How (Not) to Win a Poll

I got this leaflet for Lin Pei-hsiang 林沛祥 in my mailbox today. Lin is vying for the KMT nomination in Keelung City. On the right side of the fold, Lin teaches voters how to respond if they are sampled in the KMT’s polls. I don’t think this ad hurt Lin, but I don’t think it helped him very much either.


The most basic problem is that Lin put this leaflet in my mailbox – and presumably everyone else’s mailbox as well. Lin wants to teach his supporters how to effectively answer the survey questions, but instead he is teaching everyone, including people who support his opponents. I have heard of a candidate who did much better than expected in the polls by activating a local organizational network. Instead of teaching everyone, the campaign only taught a few trusted people in each neighborhood. Each of those people was then supposed to find a couple dozen people who they could count on to support the candidate and teach them how to answer correctly. In this way, that candidate only educated supporters and helped to maximize the effects of their support. Lin isn’t doing that. He doesn’t have a good enough network (unlikely), doesn’t know who his voters are (perhaps), or doesn’t understand how to game the system (probably).

The orange characters in the purple section instruct voters to tell the interviewer, “I only support Lin Pei-hsiang.” I’m not sure why they want the respondent to say this (and everyone seems to want it), since that doesn’t count as an answer to any specific question and, as the line right below it points out, they have to answer all nine questions for the interview to be considered a valid case.

Ok, let’s look at the nine questions. First, “Which of Keelung’s administrative districts is your household registration in?” Hold on, that’s not the first question! The first questions in KMT polls are, “How many eligible voters are there in your household? How many are male, and how many are female?” The KMT probably didn’t tell Lin about these questions because they are considered pre-questions, not part of the regular survey. The purpose is to do in-house sampling. Depending on how many eligible voters there are, the interviewer will ask to speak to the oldest male, the second youngest female, or whoever else the sampling table indicates. In the KMT protocol, if the first person is unavailable, they can substitute one other person (according to the table). If that person is also unavailable, the interviewer thanks the person and hangs up. This process is intended to ensure that there is no bias in the sample due to particular types of people always being at home or always answering the phone. From a candidate’s point of view, the difficult part is for the polling house to sample one of their supporters. It would be terrible if one of their loyal supporters answered the phone, and then was rejected because the interviewer demanded to talk to someone else who wasn’t at home. Don’t laugh, this happens all the time. If there are three eligible voters, one is not home from work yet and another is taking a shower, the interviewer will thank the person (who was ready to express support) and hang up. However, in-house sampling is an opportunity for organized campaigns. If you can teach (only) your supporters how to answer, tell them to only answer “one” or “two,” regardless of how many people actually live there. This ensures that either the person answering the phone or another person who is present will have an opportunity to respond to the survey. If your campaign can teach supporters this trick and other campaigns cannot, you have an advantage. (Of course, you have to teach this trick off the record. It wouldn’t look good if you passed out a leaflet telling voters to lie to your party’s pollsters. This is another way in which organization on the ground can help you in a poll.)

Ok, question 1. Which district do you live in? The leaflet doesn’t teach them how to answer. Since Keelung is a single district, it really doesn’t matter. The only wrong answer is that your household registration is actually in New Taipei City. For goodness sakes, don’t tell them that. (If you are not an eligible citizen, don’t tell them that either!)

Question 2 starts the real meaty part of the survey. “According to news reports, the possible legislative candidates for Keelung City include the KMT’s Lin Pei-hsiang, DPP’s Tsai Shih 蔡適, and the PFP’s Liu Wen-hsiung 劉文雄. Among these three, which would you vote for?” The leaflet tells voters to respond, “the KMT’s Lin Pei-hsiang!”

[Aside: I wonder if any voters who read this leaflet and are actually interviewed will be confused when the interviewer mentions the DPP’s “Tsai Shih-ying 蔡適應.” It’s not a good sign that Lin Pei-hsiang’s campaign doesn’t know the correct name of their expected principle opponent. Also, the DPP hasn’t nominated Tsai yet. I guess the KMT isn’t giving any respect to the other two DPP aspirants.]

Question 3-6 ask the same question, but substitute the other four KMT candidates in the race. Respondents are instructed to answer, “I don’t support any of the them.” This is a fine answer for Lin. However, if you really wanted to support Lin, you would tell the interviewer you supported the DPP candidate in these four matchups. That doubles the effect of your support for Lin. However, many KMT identifiers can’t bring themselves to express support for another party, so it may be best not to ask them for too much. Also, Lin wouldn’t want to tell voters to do this in a publicly distributed leaflet. There might be repercussions. Again, an unobservable network teaching supporters one-on-one is the way to go.

Questions 2-6 count for 85% of the result. Question 7, the internal party matchup, is the remaining 15%. “According to news reports, in the KMT there are several people who wish to run for legislator, including Lin Pei-hsiang, Ho Sheng-lung 何聖隆, Han Liang-chi 韓良圻, Lu Mei-ling 呂美玲, and Yang Shih-cheng 楊石城. Among these five, who do you support to become our legislator?” Of course, voters are asked to respond, “Lin Pei-hsiang.”

Questions 8 and 9 are demographics (age and education level). No instructions are given. I’m not sure if the KMT uses these questions to do weighting. If they do, I suppose you could really maximize your influence by claiming to be a 23 year old with a post-graduate degree. However, the more people who do that, the less influence each one has. It’s probably not worth it for a campaign to try to organize its supporters to lie on the demographics questions.

I don’t think this ad is helping Lin Pei-hsiang do better in the polls. I don’t think it is hurting him either. The only really helpful pieces of advice are 1) the reminder that the respondent has to complete all nine questions, and 2) the instruction to not support KMT candidates in the other four matchups. I supposed it is possible that more of Lin Pei-hsiang’s supporters will read this leaflet (and supporters of other candidates will spit on it and throw it away in disgust), so he might get a mild advantage. However, any advantage will be very small.

I do expect Lin Pei-hsiang to win the KMT nomination, and he will probably also win the seat in January. There are advantages to being the son of a mayor (1989-1997) and a longtime legislator (1995-present) in a blue city.

8 Responses to “Voter Education: How (Not) to Win a Poll”

  1. ジェームス (@jmstwn) Says:

    I see he’s advertising that he has the endorsement of John Chiang. Makes me wonder how much they will feature his son Chiang Wan-an as the golden boy of party renewal this campaign.

  2. ジェームス (@jmstwn) Says:

    Loved this post btw. Polls are still landline only, right? I’m cellphone-only, and I’m bummed I never hear from pollsters but car insurance salesmen still find me every month.

  3. R Says:

    If DPP decides to support 劉文雄 instead of nominating their own candidate, what do you think is the chance for him to prevail?

    • frozengarlic Says:

      It’s not that easy. If the DPP simply steps aside, many (most?) of their voters will not simply transfer support to Liu. They might sit it out or vote for one of the minor party candidates. You might also see a DPP local person run anyway. For this kind of bargain to work well, there has to be a grand bargain between the two national parties. The PFP would have to commit to supporting the DPP in the presidential election, and the DPP would have to commit to throwing its full weight behind a few PFP legislative candidates. Further, you would probably have to have some solid polling evidence that PFP identifiers and DPP identifiers in the relevant districts actually intended to honor the bargain. Otherwise, the whole thing would simply unravel.

      I’m also not sold on Liu as a credible candidate in Keelung. The last legislative race he ran there was in 2001. How much local support does he have left after 15 years of neglect?

      • R Says:

        Thanks for the answer,

        I’m aware it won’t be an simple nor easy endeavor to pursue. However looking at how the DPP have delayed nominating till now, & at how there seems to be a real possibility that at least DPP is considering supporting Huang Shan-shan, I can’t help but wonder about the possibility in Keelung too. If it develop into a three way race though it might be real exciting to keep an eye on.

      • Pat Says:

        I doubt the DPP would consider yielding Keelung after breaking the 50% barrier in the mayoral election last year. The race will still be tough for them but they have a better chance than ever.

    • frozengarlic Says:

      Also consider this from the local politicians’ point of view. A DPP city councilor in Keelung might never get a better chance to move up to the legislator. Every other year, the DPP candidate is basically cannon fodder. This might be the only realistic chance they ever get. Should they simply step aside? In Nangang/Neihu, Kao Chia-yu might be thinking along similar lines. If she can get into the legislature, even for only one term, it opens up a whole new set of career possibilities. If she can burnish her reputation there, she might move over to the party list, go into an executive position, or even run for mayor. If she stays in the city council, she could get stuck there (like Huang Shan-shan). This election might be the key to her entire political career. How can she afford to pass it up?

Leave a Reply

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

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

Google+ photo

You are commenting using your Google+ 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 )


Connecting to %s

%d bloggers like this: