Measuring Human Rights Abuse from Access to Information Requests

• Author: Sarah A. V. Ellington,Aaron Erlich,Benjamin E. Bagozzi,Brian Palmer-Rubin,Daniel Berliner
• Date: 01 February 2022
• DOI: 10.1177/00220027211035553
• Published date: 01 February 2022
• Subject Matter: Data Set Feature

2022, Vol. 66(2) 357 –384

Measuring Human Rights

Abuse from Access to

Information Requests

Sarah A. V. Ellington

1

, Benjamin E. Bagozzi

1

,

Daniel Berliner

2

, Brian Palmer-Rubin

3

,

and Aaron Erlich

4

Abstract

Existing measures of human rights abuses are often only available at the country-year

level. Several more fine-grained measures exhibit spatio-temporal inaccuracies or

reporting biases due to the primary sources upon which they rely. To address these

challenges, and to increase the diversity of available human rights measures more

generally, this study provides the first quantitative effort to measure human rights

abuses from textual records of citizen-government interactions. Using a dataset

encompassing over 1.5 million access-to-information (ATI) requests made to the

Mexican federal government from June 2003 onward, supervised classification is

used to identify the subset of these requests that pertain to human rights abuses of

various types. The results from this supervised machine learning exercise are vali-

dated against (i) gold standard ATI requests pertaining to past human rights abuses in

Mexico and (ii) several accepted external measures of sub-national and sub-annual

human rights abuses. In doing so, we demonstrate that the measurement of human

rights abuses from citizen-submitted ATI request texts can provide measures of

human rights abuse that exhibit both high validity and notable spatio-temporal

specificity, relative to existent human rights datasets and variables.

1

Department of Political Science and International Relations, University of Delaware, Newark, DE, USA

2

Department of Government, London School of Economics, United Kingdom

3

Department of Political Science, Marquette University, Milwaukee, WI, USA

4

Department of Political Science, McGill University, Montreal, Quebec, Canada

Corresponding Author:

Benjamin E. Bagozzi, Department of Political Science and International Relations, University of Delaware,

405 Smith Hall, 18 Amstel Ave., Newark, DE 19716, USA.

Email: bagozzib@udel.edu

Journal of Conflict Resolution

ªThe Author(s) 2021

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/00220027211035553

journals.sagepub.com/home/jcr

Data Set Feature

358 Journal of Conflict Resolution 66(2)

Keywords

human rights, machine learning, measurement, Mexico, event data

Introduction

The measurement of human rights abuse is of centr al relevance to the study of

human rights. Three of the most well-established country-year measures of human

rights abuses produced this decade—the CIRI Human Rights Dataset (Cingranelli

and Richards 2010; Cingranelli, Richards, and Clay 2014), the Political Terror Scale

(PTS; Wood and Gibney 2010), and the Latent Human Rights Protection Scores

(Fariss 2014)—have now collectively received over 2,000 citations.

1

Adding addi-

tional nuance to these datasets necessitates measurements of human rights abuses at

spatio-temporal scales that are more precise than the country-year unit (Cordell et al.

2019b). Such data would allow for more fine-grained tests of the determinants of

human rights abuses, and for further synergy with theories and models of political

violence—which have increasingly shifted towards sub-national and sub-annual

data over the past decade (Cederman and Gleditsch 2009; Raleigh et al. 2010;

Gleditsch, Metternich, and Ruggeri 2014). These advancements would also offer

scholars and advocacy groups (i) the ability to better detect and preempt human

rights abuses before they spread and (ii) improved understandings of the microfoun-

dations of human rights abuses.

In light of these evolving measurement needs, this article considers the use of

textual records of citizen-government interactions for the development of new

fine-grained spatio-temporal data on human rights abuses. Such information and

communications technology (ICT)-enabled platforms include citizen reporting

initiatives, complaint mechanisms, official social media accounts, and our focus

here: access-to-information (ATI) requests. Across the world, country-specific

ICT platforms increasingly make large-scale textual records of past

citizen-government interactions available online, thus offering researchers and

practitioners new opportunities to measure real-world outcomes at a

fine-grained level.

We specifically consider data from one ATI regime for which we have access to

comprehensive records of every single individual request for government informa-

tion: the case of the Mexican federal government. Following Mexico’s landmark

2002 ATI law, all individual ATI requests filed with Mexican federal government

agencies are publicly available. Additional requests made to other federal branches

of government and constitutionally autonomous bodies were added to this publicly

accessible system in 2016. Each individual Mexican ATI request includes the textual

description of the information that a citizen or organization seeks, supplemental

attachments, textual entries that contextualize the requested information, the reques-

ter’s municipality, the date of the request, the government entity to which the request

is directed, and information the Mexican government’s response to each request.

2

2Journal of Conflict Resolution XX(X)