Self-Censorship Under Law A Case Study of the Hong Kong National Security Law Mona Wang

2025-05-03 0 0 185.4KB 10 页 10玖币

侵权投诉

Self-Censorship Under Law:

A Case Study of the Hong Kong National Security Law

Mona Wang

Princeton University

Jonathan Mayer

Princeton University

Abstract

We study how legislation that restricts speech can in-

duce online self-censorship and alter online discourse,

using the recent Hong Kong national security law as a

case study. We collect a dataset of 7 million historical

Tweets from Hong Kong users, supplemented with his-

torical snapshots of Tweet streams collected by other

researchers. We compare online activity before and af-

ter enactment of the national security law, and we nd

that Hong Kong users demonstrate two types of self-

censorship. First, Hong Kong users are more likely than

a control group, sampled randomly from historical snap-

shots of Tweet streams, to remove past online activ-

ity. Specically, Hong Kong users are over a third more

likely than the control group to delete or restrict their

account and over twice as likely to delete past posts.

Second, we nd that Hong Kong users post less often

about politically sensitive topics that have been cen-

sored on social media in mainland China. This trend

continues to increase.

1 Introduction

Empirical research about censorship of online speech fo-

cuses on measuring network- and platform-level content

restrictions. Governments do often censor online speech

by outright preventing access to content, but there is an-

other important form of censorship that has received far

less quantitative study: self-censorship, when a govern-

ment chills online speech by imposing legal (and some-

times extralegal) risks. Despite a wealth of qualitative

research, media reporting, and public writing about self-

censorship, there have been relatively fewer large-scale

quantitative case studies that demonstrate the eect of

speech law on online self-censorship at scale.

On June 30, 2020, the Chinese National People’s

Congress enacted a “national security law”(NSL) with

direct application to Hong Kong [27]. The NSL en-

tered into force the same day. The new law criminalizes

speech that the government deems “seditious”or “se-

cessionist”in nature, terms which the law implements

with broad and ambiguous prohibitions.

Anecdotally, individuals and organizations in Hong

Kong have recently curtailed their own activities and

censored their own speech both oine and online, espe-

cially after the national security law (NSL) entered into

force. Users on LIHKG, a social media site popular in

Hong Kong, suggested that Hong Kongers delete social

media accounts or past social media activity that could

incriminate them under the NSL [7]. Since the NSL’s

enactment, Hong Kong authorities have arrested people

on the basis of their online activity, and local lawmakers

have proposed additional bills to outlaw specic forms

of online speech (e.g., “Publishing an image of a deled

national ag on Facebook”) [6,9].

Our work aims to demonstrate the impact that newly

enacted liability can have on online speech, using Hong

Kong’s national security law as a case study. We quan-

tify at scale the self-censorship exhibited by social me-

dia users after the national security law passed. We

use Twitter datasets because, according to social media

market research, about 29% of Hong Kongers aged 19-

64 used Twitter in 2021 [21]. Our work seeks to answer

two research questions.

RQ1: Comparing social media activity by Hong

Kong users before and after enactment of the national

security law, how often do Hong Kong users delete posts

or accounts, and how does the frequency compare to

other groups of users?

RQ2: Comparing social media activity by Hong

Kong users before and after enactment of the national

security law, how has the amount of discussion of

sensitive political topics possibly covered by the NSL

changed relative to other topics, and how does this dis-

cussion compare to a control group?

To answer these research questions, we curate several

datasets of Tweets and Twitter users. First, we compile

a dataset of archived social media data from before the

NSL entered into force. This historical dataset includes

posts and accounts that may have subsequently been

deleted or made private. The archived data consists of

about 2 million Tweets from various user populations

during 2019. Second, we compile datasets of currently

available Tweets and Twitter users from before and af-

ter enactment of the NSL. These datasets contain over

7 million Tweets from Hong Kong users and 8 million

Tweets from a set of control users.

In our analysis to answer RQ1, we nd that Hong

Kong users are over a third more likely than a control

sample to protect their accounts and over twice as likely

to delete past Tweets than control Twitter users.

To address RQ2, we additionally curate a dataset of

Tweet keywords that were common among Hong Kong

users before the NSL and that are associated with polit-

ically sensitive topics that are censored on social media

platforms in mainland China. We analyze the relative

frequency of Tweets containing politically sensitive key-

words over time for Hong Kong users and for a control

group. We nd that Hong Kong users continue to speak

less online about politically sensitive topics.

Our case study presents large-scale quantitative evi-

dence that aggressive legislation and policy can quickly

and starkly alter the nature of online political discourse.

2 Background

In this section we present prior research measuring self-

censorship in online discourse, and we oer background

for our Hong Kong case study.

2.1 Measuring self-censorship in online

discourse

We dene self-censorship, or the chilling eect, consis-

tent with prior scholarship: when an individual with-

holds or falsies discourse for fear of repercussion [20].

There is a vast literature on measuring online political

discourse [11,13,25]. There is also a large body of quali-

tative research, especially in law, public policy, and poli-

tics, about self-censorship and chilling eects [14,19,20].

The media also often reports on this phenomenon, of-

ten from anecdotal evidence or hypotheses by policy-

makers [16–18]. There is, however, very little large-

scale empirical research on changes in online political

discourse that are attributable to self-censorship.

Past research has shown that measurable dierences

can surface around discrete events that increase the per-

ception of online surveillance. In the most similar prior

work, Tanash et al. quantied the change in Tweet-

ing behavior by Turkish users specically after the 2016

attempted coup in Turkey [23]. The Turkish govern-

ment subsequently arrested thousands of people that it

blamed for plotting the coup, with little due process.

Many of these arrests resulted from investigations into

social media activity, solely on the basis of individu-

als’ online speech and actions. Notably, Tanash et al.

measured both a surge in retroactively deleted tweets

by Turkish users and a signicant decrease in certain

politically sensitive tweets from Turkish accounts.

2.2 The Hong Kong national security

law

The new national security law for Hong Kong creates

penalties for people who participate in secession, sub-

version of the governments of mainland China or Hong

Kong, terrorist activities, or collusion with a foreign

country to endanger national security. In addition to

having a vague and sweeping scope, the law extends

beyond Hong Kong: Article 38 establishes liability for

oenses that occur “outside the region by a person who

is not a permanent resident of the region”[27].

In the six months after the NSL entered into force,

Hong Kong law enforcement arrested at least 100 indi-

viduals on the basis of the new law [26]. At least 24

of the arrests involved charges related to “seditious”or

“secessionist”speech. The arrestees included legislators,

protestors, student activists, journalists, and an Amer-

ican human rights lawyer. Journalists in Hong Kong

have described their fear of declining press freedoms and

increased self-censorship in the media [3]. Because en-

forcement of the NSL has already targeted online polit-

ical speech, Hong Kongers may have a strong incentive

to self-censor their online social media activity.

3 Methodology

In this section, we describe how we collect data to an-

swer our two research questions. For each, we curate

several large datasets, and we perform various analyses

on the data.

To curate these datasets, we combine various sources

of Twitter data both from archives and from Twitter’s

Full-Archive Search API. We then augment the data

with additional live data from the Twitter API. Next,

we lter and curate these large data sources into smaller

datasets, which we use for analysis. We enumerate our

datasets in Table 1.

Figure 1illustrates the data collection process for our

study. The code for the data collection and analysis can

be found at https://github.com/citp/hk-twitter.

3.1 Post and account deletion

RQ1: Comparing social media activity by Hong Kong

users before and after enactment of the national secu-

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

Self-CensorshipUnderLaw:ACaseStudyoftheHongKongNationalSecurityLawMonaWangPrincetonUniversityJonathanMayerPrincetonUniversityAbstractWestudyhowlegislationthatrestrictsspeechcanin-duceonlineself-censorshipandalteronlinediscourse,usingtherecentHongKongnationalsecuritylawasacasestudy.Wecollectadataseto...

展开>> 收起<<

Self-Censorship Under Law A Case Study of the Hong Kong National Security Law Mona Wang.pdf

共10页,预览2页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Self-Censorship Under Law A Case Study of the Hong Kong National Security Law Mona Wang

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: