Quote:
Originally Posted by wao
|
Please, wao, a little mercy for us Japanese illiterates in the audience.
Part of the problem is that the times listed at Mahou didn't match up with the ratings data. Mahou reports Bartender running on Sundays at 1:45 am, but in the ratings data it appears at 26:25 on Saturday. It wasn't obvious to me what a time like 26:25 on Saturday means; is it 2:25 am on Saturday, or 2:25 am on Sunday? In either case, it wasn't listed at 25:45 which I presume is what 1:45 am would translate to in this (somewhat-wacky) time scheme.
I really did try, but not having a reading knowlege of Japanese, and the fact that the sources didn't match, made this a more difficult task. Also the relationship between the network codes and the names of the networks isn't obvious either; for example, Fuji somehow becomes CX.
According to Mahou, Saiunkoku Monogatari aired on Saturday mornings at 9:00 am (!) on NHK/BS2 beginning on April 8, 2006. Only one show is listed in that time slot in the ratings data, one that translates to "Super star fleet [seiza] X" according to Google, which was shown on channel "TX" and had ratings for weeks prior to 4/8/06. There are other entries for NHK shows, with the code "NHK," but not one for SaiMono, I guess.
On ratings methods:
Back in the early to mid-80's I worked on a foundation-sponsored project to develop an alternative to "head-count" ratings based on viewers' opinions about the show they were watching, and their behavior while it was on. This is probably what Nielsen means by "popularity" in the material rooboy cites. These so-called "qualitative" ratings are used in many other countries, including Japan, but never caught on in the US. It's use in Japan is probably a reason for the ratings != popularity refrain to which Kaoru Chujo refers.
We could actually show that people were more likely to watch ads in programs to which they paid higher levels of attention, but advertisers weren't interested. Qualitative data on programs is still collected in the US by programmers, networks, and advertisers, but usually through broad surveys like roo suggests, not through a ratings instrument.
What did become the standard for "qualitative" information in the US, if you can call it that, is better demographics, especially via the "people meter." Ratings were traditionally collected at the household level with meters attached to every TV in the household. People meters include separate buttons for each household member who are supposed to "log in" when they start viewing and "log out" when they leave. Obviously this procedure raises considerable questions about the accuracy of the data, since it depends on the active cooperation of the viewers in the sampled households. There were a number of efforts to develop "passive" people meters that somehow would be able to sense which people were present in the room without active cooperation, but they always seemed to founder on tasks like differentiating between small children and large dogs! Nowadays we could probably just
embed RFID tags in each person and let the meters detect those.
My guess, and it's only a guess, is that the numbers reported on the Geocities site are household figures and thus include DVR recordings as well as real humans. Making sure that ratings meters collect data from all viewing sources, including DVDs, games, etc., has been a high priority in the television research community for many years now.
Has anyone seen comparable figures for shows that run in [adult swim] in the US? What does something like Inuyasha or FLCL draw?