What information do citizens want? Evidence from one million information requests in Mexico

Abstract

While scholars have emphasized the importance of information for accountability, little research has addressed the demand for government information by real citizens. We study the totality of information requests filed with Mexican federal government agencies from 2003 to 2015, over 1 million requests in all. We use unsupervised methods to categorize requests, revealing the diversity of topics including environment, security, budgets, and government procurement and employees. While many topics have clear public accountability-seeking purposes, others are focused on more private, micro-political goals. Analysis over time and across states reveals linkage between information demand and issues of public interest such as environmental impacts and criminal violence. Our results demonstrate that, given functioning access-to-information institutions, citizens in a transitional democracy really do demand information relevant to public accountability.

When does open government shut? Predicting government responses to citizen information requests

“…Because omitting these would yield an unrepresentative sample, we separately downloaded each attachment, digitized the text, and appended it onto the main request text field. 9 Consistent with past research (e.g., Bagozzi, Berliner and Almquist, 2016;Berliner, Bagozzi and Palmer-Rubin, 2018), we then truncated all remaining texts from the thousandth string onwards. This created our primary corpus, which was further preprocessed using standard steps for the automated analysis of political texts.…”

“…Topic models have been shown to be highly valid for the discovery of latent thematic content within information request texts (Berliner, Bagozzi and Palmer-Rubin, 2018;Berliner et al, 2020). As such, the present article evaluates the utility of supervised latent Dirichlet allocation (sLDA) models (Blei and Mcauliffe, 2003) for the prediction of bureaucratic (non)responsiveness to these same request texts.…”

“…2016; Berliner et al . 2018), we then truncated all remaining texts from the thousandth string onward. This created our primary corpus, which was further preprocessed using standard steps for the automated analysis of political texts.…”

“…Topic models have been shown to be highly valid for the discovery of latent thematic content within information request texts (Berliner et al . 2018; Berliner et al . 2015).…”

Media Attention and Bureaucratic Responsiveness

“…Information requests offer a tool for citizens to engage with every facet of an agency's operations. Research has shown both that the volume and focus of information requests change in line with key events (Berliner, Bagozzi, and Palmer-Rubin 2018) and that agency personnel exercise discretion in responding (or not) to information requests, motivated by partisan electoral goals as well as personal career incentives (Berliner et al 2020). Future research should expand our approach to observe whether these results replicate to other modes of citizen-state interaction that are less politically relevant.…”