Asahi, Yomiuri, and the Kyodo News wire service ran Oct. 1-2 polls right after the media issued reports on the new Noda cabinet. The following are the results for three key questions, with the previous September poll numbers in parenthesis. The only consensus there is that the LDP is up, while DPJ numbers are scattered. What’s truly anomalous, though, are support and voting intent numbers for Osaka Mayor Toru Hashimoto’s Japan Restoration Party (JRP); they look extremely out of kilter across the board. What’s going on?
Here’s what I think. The numbers in a hypothetical poll that perfectly reflects national demographics would be closer to the Kyodo numbers, but the JRP needs to get its act quickly, in particular putting an end to what looks like a power struggle between Hashimoto and its Diet members, who defected from the LDP, DPJ, and Your Party. Otherwise it will see its numbers slipping to Asahi levels. Let me explain.
Polls tend to reflect the views of the media outlet that is taking the polls. Thus, the LDP, nuclear power and the like tend to do better on Yomiuri polls than on Asahi polls. Although the polling companies conduct their surveys by random digit dialing, they first do declare on whose behalf they are taking the polls. Personal experience and common sense tell me that we are more inclined to spend our time answering pointed questions from strangers when
theywe are being polled on behalf
of beneficiaries that theywe have some familiarity with sympathy towards.
Thus, a newspaper poll should be disproportionately weighted towards its
subscribers, who are likely to have been heavily influenced over the years by
the editorial slant that permeates the news items*. No wire service elicits
that kind of brand loyalty. From that perspective alone, the Kyodo wire service
poll should provide less biased results.
* The board of editorial writers and the editorial board are all promoted from the ranks of the reporters, giving Japanese newspapers and wire services a seamless continuity from the correspondents to the editors to much of management.
There’s more. The national newspapers dominate the metropolitan Tokyo-Yokohama and metropolitan Osaka-Kobe areas and their immediate environs (respectively: Tokyo, Kanagawa, Chiba, Saitama, Ibaragi; and Wakayama, Nara, Hyogo). In prefectures elsewhere, local newspapers account for roughly one-half to four-fifth of total subscriptions.* This means that roughly one-half to four-fifths of newspaper subscribers outside of the Tokyo and Osaka neighborhoods will be less inclined to respond to any newspaper poll. These local-newspaper subscribers should be getting the bulk of their printed national news from the wire services. If, as I assume, they have less brand-loyalty to wire services than to newspapers, they are unlikely to be overrepresented in the prefectural and national Kyodo numbers. In short, the Kyodo numbers should be relatively free of geographical bias as well.
* A quick 2004 overview, a little old but shouldn’t be far off even now.
Why, then, does the JRP do better in every which way in the Kyodo polls? The local newspapers by definition carry more local news. They also have fewer pages. I suspect that much of the commotion around the somewhat confusing launch of the JRP and subsequent dissonance have been given short shrift in the local papers, which enabled it to escape much of the resultant fallout. Still, it appears to me that the JRP needs to put the commotion behind itself and project a confident, united front going forward, or else the lack of positive reports, coupled with whatever negative reports come in through print and TV coverage will further erode support.
Note also that voter preference for the JRP in the conservative Yomiuri poll is still pretty high in contrast to the more general support. It’s obvious that support for the JRP is still highly transient and the next couple of months, when the JRP lines up candidates and sets up a definitive campaign platform, will be crucial to the JRP’s long-term viability.
Much of what I’ve written is speculative, but it’s the best that I can do, given the time and resources available to me. Hopefully, polling experts can convince media outlets to conduct a more in-depth analysis of these biases and their effects.
Q. Which party do you support?
Asahi: LDP 21 (15); DPJ 14 (16); JRP 2 (3)
Kyodo News: LDP 30.4 (19.3); DPJ 12.3 (12.9); JRP 10.7 (?)
Yomiuri: LDP 28 (21); DPJ 18 (15); JRP 2 (2)
Q. Which party do you want to vote for on the lower house proportional district ballot?
Asahi: LDP 30 (23); DPJ 17 (15); JRP 4 (5)
Kyodo News: LDP ??.? (22.2); DPJ 12.3 (12.4); JRP 13.9 (17.6)
Yomiuri: LDP 36 (31); DPJ 18 (14); JRP 13 (16)
Q. Do you or do you not support the Noda cabinet?
Asahi: Yes 23 (25); No 56 (53)
Kyodo News: Yes 29.2 (26.3); No 55.3 (59.4)
Yomiuri: Yes 34 (27); No 56 (63)
Random digital dialing survey; fixed line-only, no cellphones.
Previous poll numbers in parenthesis: Kyodo News 09.01-02, Asahi 09.08-09, Yomiuri 09.15-17.September numbers for JRP: Osaka Ishin-no-Kai for Kyodo News and Asahi.