East Asia and English language speakers: a population estimation through existing random sampling surveys

ABSTRACT

Many applied linguists have attempted to quantify the number of English-speaking populations, both globally and in specific countries. However, thus far, those in East Asia have not been estimated through a robust technique. This paucity of estimation can be partly attributed to the lack of censuses or large-scale sociolinguistic surveys; nonetheless, other secondary data can be used for this purpose. This study analyses two nationally representative surveys conducted in China, Japan, South Korea, and Taiwan – the East Asian Social Survey (conducted in 2008 and 2018) and AsiaBarometer (2003–2008) – and estimates their English-speaking population. The results indicate that English speakers (those who evaluated themselves as having a fair or high level of English proficiency) account for approximately 10% of the population in China and Japan and approximately 20% in South Korea and Taiwan. Furthermore, English proficiency varied according to generation and educational level in all four jurisdictions. Based on these findings, this article discusses cross-national variations of English-speaking populations in the four jurisdictions, the absolute number of speakers, its implications, and possible self-report biases and their assessment.

KEYWORDS:

• English-speaking population
• East Asia
• statistical estimation
• nationally representative sample
• secondary analysis
• generational effect

Acknowledgments

The author would like to express sincere gratitude to the East Asia Social Survey Data Archive (https://www.eassda.org/) for providing the datasets used in this research. The author also would like to thank the AsiaBarometer Project for providing the dataset. The dataset detail is as follows: Inoguchi, Takashi and Shigeto Sonoda, AsiaBarometer Integrated Dataset, [computer file]. AsiaBarometer Project (http://www.asiabarometer.org/) [producer and distributor], downloaded on 15 January 2022. AsiaBarometer is a registered trademark of Professor Takashi Inoguchi, President of University of Niigata Prefecture, Japan, Director of the AsiaBarometer Project.

Disclosure statement

No potential conflict of interest was reported by the author.

Data availability statement

The data that support the findings of this study are available in the repositories of the EASS (http://eassda.org/) and AsiaBarometer (https://ricas.ioc.u-tokyo.ac.jp/aasplatform/download/using.html). For more details, see note 6.

Notes

1. Although Bolton and Bacon-Shone’s (2020) estimation of English speakers of China and Japan is not entirely guesswork, they do not refer to nationally representative surveys.

2. Many studies have demonstrated such a bias. For example, Terasawa (2021) conducted a statistical comparison between a nationally representative survey of Japan and a survey using research company’s web panels, revealing that the latter contained a much larger number of respondents with English proficiency than the former.

3. Subjective correction is a method by which figures obtained from data are corrected based on the analyst’s subjective interpretation (e.g. ‘I find that this survey has a larger number of highly educated people than the population, so I should correct this estimate to a smaller value.’). Statistical correction calculates a survey’s degree of possible bias in using statistical modelling based on other valid datasets (e.g. censuses) and adjusts an estimate according to this bias score.

4. There are several nationwide surveys with random sampling that investigated citizens’ English proficiency in the 2010s. For example, the Japanese General Social Surveys and the China General Social Survey included items on English proficiency, which various researchers have analysed (Terasawa, 2018; Zhang & Lien, 2020).

5. Note that AsiaBarometer is not the same as Asian Barometer. The latter is led by the Hu Fu Center for East Asia Democratic Studies.

6. The individual dataset of the EASS can be downloaded from http://eassda.org/. The dataset of AsiaBarometer can be accessed via https://ricas.ioc.u-tokyo.ac.jp/aasplatform/download/using.html or the AsiaBarometer Project, University of Niigata Prefecture, Tokyo Satellite, Koyosha KS Building 9th Floor, 1-17-8 Nishikata, Bunkyo-ku, Tokyo 113–0024, Japan ([email protected]).

7. For statistical analysis, the author used R 4.1.0 and its default packages.

8. A questionnaire translated from English into local languages was presented to respondents. However, it should be noted that the translation of the three items was not always a literal translation. Items in local languages were as follows. For China, (A) 阅读英文报纸中的短文, (B) 用英语与人对话聊天, (C) 用英文写封信. For Japan, (A) 英字新聞の短い記事を読む, (B) 英語でおしゃべりする, (C) 英語で手紙を書く. For South Korea, (A) 영자신문 기사 읽기, (B) 영어권 사람과 대화하기, (C) 영어로 편지 쓰기. For Taiwan, (A) 閱讀英文報紙中的短文, (B) 用英文與人聊天, (C) 用英文寫封信.

9. The translated scale for each language is as follows. For China, 5 = 非常好, 4 = 好, 3 = 普通, 2 = 不好, 1 = 非常不好. For Japan, 5 = 非常によくできる, 4 = よくできる, 3 = 少しはできる, 2 = あまりできない, 1 = ほとんど/まったくできない. For South Korea, 5 = 매우 잘한다, 4 = 잘하는 편이다, 3 = 보통이다, 2 = 못하는 편이다, 1 = 전혀 못한다. For Taiwan, 7 = 非常好, 6 = 好, 5 = 有點好, 4 = 無所謂好不好, 3 = 有點不好,2 = 不好, 1 = 非常不好.

10. The original passage written in local languages is almost identical. Only China’s option (B) in the 2008 version (用英文与人聊天) differs slightly from the 2018 version.

11. Notably, comparison between the 2008 and 2018 percentages by region (except for Taiwan) and English language skill revealed some statistically significant differences (2008 < 2018). However, Cramer’s V coefficients for the degree of difference (effect size) were generally small, with the largest being V = 0.088. This indicates that statistical significance was mainly due to this survey’s large sample size, and it is difficult to conclude that any substantial changes had occurred during these 10 years.

12. China’s sharp decline in the graduate school graduates can be attributed to the small number of cases (China’s datasets included only nine respondents with this educational background).

13. This exclusion was indispensable because the 2003 and 2004 surveys investigated those aged 20–59 years, but the 2005–2008 surveys included those aged 20–69 years.

14. An academic database search (e.g. Google Scholar) using the phrase ‘English frenzy’ seems to produce more articles examining South Korea and Taiwan, than China and Japan.

Additional information

Funding

This work was supported by JSPS KAKENHI [Grant Number 22K00726], [Grant Number18K12480].