Harnessing the power of data science to plan for human rights impact

By: Zehra Mirza, Director of Impact and Learning, Amnesty International USA

In an era of information misuse, Amnesty International USA is using data science to advance human rights impact.

Monitoring and evaluation practices establish roadmaps for organizations to achieve change. By analyzing traditional structured data, evaluations distinguish impact by understanding the levers that bring about a certain outcome. These practices, however, can be repurposed to use big data as a means for actionable planning. While traditional forms of structured data can be easily recorded and evaluated, big unstructured data is messier and less well-understood, at times resulting in its exploitation, such as the case of Cambridge Analytica. Although, when big data is used appropriately, it can guide the development of an effective human rights strategy.

When we think about data-based strategies in the human rights sector, we often think about their methodological constraints. These strategies can be too rigid, time-consuming, and impractical to keep up with the fast-paced nature of activism. Windows of opportunity are unpredictable and require agility, leading advocates to rely on their instincts rather than a systematic assessment of their context.

Amnesty is merging the power of data science with campaign planning to make informed decisions in a way that is relevant and meaningful to its advocates. As part of its most recent crisis campaign to end the use of excessive force by police, Amnesty leveraged big data to provide recommendations for campaign messaging. Through frame analysis, the organization crafted nuanced and tailored messaging for key legislative actors and audiences to drive human rights impact. Staff analyzed public information by mining key terms such as police brutality, chokeholds, George Floyd, and Black Lives Matter from social media posts to create rapid assessments on observable trends in language and sentiment. These analyses helped prepare Amnesty’s Lobby Day activists for their meetings with members of Congress and are now shaping the way Amnesty develops op-eds to influence other target groups.

Image for post
Image for post
Image for post
Image for post
The images above are an example of the text mining analyses of a key legislator’s social media posts. The findings showed what terminology the legislator uses most when describing excessive force by police and the sentiment of that text.

Amnesty’s rapid assessments also include scanning key terms to see what cities in the United States are discussing deadly force and police accountability most and how they are framing the discussion. Understanding this geographic variability in language and sentiment of a human rights issue, enables staff to develop content using select terms that are yielding the greatest engagement among their target audiences. For example, a rapid assessment of select media outlets in Utah, South Carolina, and Kentucky showed what language is trending in those areas and informed how an op-ed could be best written to generate action from target constituencies.

Image for post
Image for post
Image for post
Image for post
*Grey shaded states indicate very few searches

Monitoring and evaluation methods based on big data are relatively new and are being tested in the field.

Amnesty is keen on uplifting them as part of its commitment to innovation. Typically, a frame analysis is used post-facto, to assess public narrative and to determine how effective a campaign has been in changing public sentiment on an issue. It can also be a valuable method, however, for advance planning in guiding the messaging direction of a campaign.

Of course, Amnesty continues to monitor and evaluate other aspects of its crisis response. Most notably, staff assess Amnesty’s membership base and its engagement with any campaign related actions such as emails or calls made to Senators. Understanding where supporters are based, and monitoring the actions they take with Amnesty, enables the organization to better mobilize in key states, where timely constituent pressure can move target legislators.

Image for post
Image for post
Image for post
Image for post
Geographic Distribution of George Floyd Petition-Takers who are NEW to Amnesty International USA

Though the above big data examples represent novelty in advocacy evaluation, traditional practices are still relevant and needed. The human rights landscape around the world is nuanced and complex, and evaluation techniques must account for context. Regardless of method, measuring any sort of change is difficult due to issues around attribution, causality, and the long-term, arbitrary trajectory of human rights impact. Despite the challenges with analyzing data and its application, Amnesty is continuing to integrate data science into its evaluation methods to guide advocacy and organizational learning.

At a time when data is exploited to exacerbate human rights crises around the world, Amnesty International USA is leveraging it as a valuable resource to trigger positive change.

Written by

We've been fighting the bad guys since 1961 - you can join us! Official Amnesty International USA profile.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store