Ethical Approaches to Closed Messaging Research: Considerations in Democratic Contexts

Written by Connie Moon Sehat and Tarunima Prabhakar, with Aleksei Kaminski

15 March 2021

From the Executive Summary

“Closed” messaging apps such as WhatsApp, Signal and Telegram have grown in reach and adoption in recent years and have transformed elections-related communications. However, understanding their impact upon larger public discussion poses a conundrum for researchers. On the one hand, widely accessible conversations of public importance exist in these spaces, but on the other, message encryption challenges existing professional ethics of access and collection. In order to study what messages are flowing through these channels, analysts and researchers must join potentially private chat groups. This raises the question: What considerations should be taken into account when conducting research into closed messaging spaces within democratic contexts, in which the individual right to privacy also prevails? 

To better understand the decisions that researchers face, this paper presents research practices taking place within these applications. It focuses on projects with election-related themes that parse the content of messages. Two investigating groups are of interest: public-interest organizations focused on human rights, democratic elections, and fact-checking that seek real-time or more immediate impact based upon their findings, and academics who want to provide longer-term, systematic analysis of the political dynamics created by these closed conversations. Sometimes these groups work in collaboration with each other as well as with other types of researchers, such as those based at technology companies.

When it comes to studies involving the systematic collection of message data, there are at least four models of research practice. These models are summarized below:

  • Baseline model: Smaller Investigations under Informed Consent
  • Model 1: Voluntary Contributions (e.g., tiplines)
  • Model 2: Joining Groups through Invitation with Public/Publicized Purpose
  • Model 3: Joining “Public” Groups Without Disclosing Research Intent
  • Model 4: Joining “Public” Groups While Disclosing Research Intent

First, we share background about the reach of closed messaging applications and the motivation for these research efforts. Summaries of the four models follow. Based on the models, the final section of the paper offers a set of key questions based on public parameters, participant safety, and researcher obligations, to help clarify the ethical tensions posed by investigations into closed message discussions:

  • Parameters: When exactly is a closed message chat “public”?
  • Parameters: Who does the data belong to?
  • Security: What is the plan for storing closed messaging data in a secure fashion?
  • Security: How might the research project affect group members and their conversations?
  • Obligations: What are the obligations for research disclosure and informed consent?
  • Obligations: When should researchers inform or report back to the groups involved?

In addition, a data collection guide is included at the end of the paper, which pairs practical steps with key ethical considerations.

Read the Full Report

<<< Back to Verifying Elections: Towards Good Practices Amid Social Media-Faciliated Misinformation