@book{oai:repository.ninjal.ac.jp:00001297,
 author = {国立国語研究所 and The National Language Research Institute},
 month = {Oct},
 note = {This is a preliminary, exploratory, systematic survey of the vocabulary used in Japanese television broadcasts. In this report, we present (1) the methodology used in the survey, (2) a list of a subset of the sample data, and (3) a statistical analysis of the subset. The complete vocabulary list and analysis are provided in the continuing volumes 2 and 3.

The purpose of this survey is to gain a fuller picture of the vocabulary used in the contemporary Japanese language and establish a fundamental methodology for surveys of vocabulary usage in television broadcasts.

The survey design follows that of previous vocabulary surveys conducted at the National Language Research Institute, and survey methodology used in sociolinguistic-phonetic research. In addition, this survey differs from survey research conducted at the NHK Broadcast Culture Research Institute, which is designed to meet specific needs of the broadcast industry because we aim to present a comprehensive view of the nature of language actually used in broadcasts.

The data for this survey consist of the complete 24-hour broadcasts of the six major television networks in Japan (seven channels: the Japan Broadcasting Corporation (NHK) general and educational stations, Nihon TV, TBS, Fuji TV, TV Asahi and TV Tokyo) recorded over a period of three months. We analyzed all uses of the Japanese language, both audio (each time a word was spoken) and visual (each time a word was continuously displayed on the screen).

Survey methodology: We divided the complete data set (population), which consisted the recorded broadcasts into five-minute segments as the basic unit for sampling. Samples were randomly chosen, keeping the quantity of data constant for each week, day of the week, time slot and channel.

Units of data: (1) We adopted the concept of “long units”, based on the definition used in Report 89. “Long units” consist of words plus adjacent particles and verbal auxiliaries. (2) We adopted the list of words specified in the Reports 21, 25 and 89 as headwords for this study. (3) Separate totals of occurrences of audio (spoken) and visual (written) uses were calculated for each word.

Scale of the survey: The complete data set consists of 15,288 broadcast hours. The cumulative total amounts to approximately 57 million audio word tokens, and 25 million visual word tokens. One 504th of the data set was sampled yielding 364 five-minute samples. The total length sampled was 30 hours and 20 minutes (96,832 seconds of programs and 12,368 seconds of commercials). Of these, 332 samples provided usable data: the audio data consist of 331 samples from television programs and 134 samples from commercials and the visual data, 301 samples from television programs and 134 samples from commercials.

Listing of data samples: We provide a list for the 364 samples (including those which occur outside of a station's broadcasting hours), including a description of each sample.

Analysis: We analyzed the data from the 364 samples exploring the relationship between various aspects of vocabulary use and extra-linguistic factors from the 12 points of view listed below.
(1) Total samples, (2) Programs versus commercials, (3) Audio (spoken) versus visual (written) vocabulary uses, (4) Program genre, (5) Television channels, (6) Broadcasting time slot, (7) Broadcasting day, (8) Program length, (9) Viewer ratings, (10) Speakers, (11) Viewer behavior, (12) Visual language media

Investigators: NAKANO Hirosi, ISII Masahiko, ÔSIMA Motoo, YAMAZAKI Makoto, ONUMA Etu, application/pdf},
 publisher = {国立国語研究所},
 title = {テレビ放送の語彙調査 1 方法・標本一覧・分析},
 year = {1995},
 yomi = {コクリツ コクゴ ケンキュウジョ}
}