Podcast: Dr. Patrick Ball on Using Statistics to Uncover Truth
Dr. Patrick Ball recently visited the Plutopia News Network podcast for a wide-ranging, inspiring conversation about his work for the Human Rights Data Analysis Group.
Patrick spoke about how he first discovered human rights work during his time in El Salvador with the Peace Brigades International. That led to his ongoing work as a statistician and computer programmer working to assess and analyze human rights violations. He also unpacked some common statistical techniques used by researchers at Human Rights Data Analysis Group, such as multiple systems estimation, which uses multiple different datasets to gain insights into the data we don’t have.
“The story of missing data is what I’ve spent my whole career thinking about,” explained Patrick. “The joke in our team is if the data were perfect, lawyers would do this with excel.”
Imperfect datasets are the norm in most instances of data involving human rights violations like state-sponsored violence, whether that’s tracking police homicides in the United States or disappearances in Colombia. As Patrick explained, it is only through statistical analysis that we can shed light on the data that is missing from the databases: deaths and violence that are unreported.
Patrick pointed out that some communities are less likely to report on state violence, often because they fear reprisal. In the United States, for example, he specifically highlighted how those living on Native American reservations, those without permanent housing, and undocumented people are all less likely to report on a police killing. At the same time, these communities are at an increased risk of experiencing state violence. This creates the perfect storm for hidden violence: As Patrick explained:
“These are also three categories of people who get killed much more often than other people. So if you put together those two observations, these are systematically under-observed categories of homicide and categories of people more likely to be killed, then you realize there’s going to be a lot of homicides that are not in anybody’s reporting.”
Statistical analysis can also be a powerful tool to use against fake news and misinformation, which often seeks to confuse public perceptions by offering untrue (and very unscientific) information. Patrick pointed out that there’s been an increase in the use of large language models to produce a lot of online content that mimics scientific information, while being based on nothing—contributing to an environment where people don’t trust anything they read. This situation makes rigorous, well-researched, reputable scientific analysis more important than ever, especially when it comes to assessing the claims of governments.
Watch the whole interview below, and please share it on Bluesky and Mastodon.
Learn more about Plutopia News Network. Follow Plutopio News Network on Bluesky.
Image: Dr. Patrick Ball at TechCrunch Disrupt in 2018. CC-BY-2.0