Identifiers of Detained Children Have Implications for Data Security and Estimation
For this investigation, I used a dataset that contains the names, alien numbers, countries of origin, birthdates, dates of entry in the detention facility, and other personal information for thousands of undocumented migrant children held in shelters in the US Midwest between 2013 and late 2018. The shelters at which the children in the dataset were detained are privately run facilities contracted by the US federal government to provide services until a suitable guardian is identified.
The children were given “alien numbers”—unique identifiers assigned by the government— that are generally sequential. Although it is not perfect, there is a positive correlation between the alien numbers and the date of entry into the detention center. This has implications for data security and estimation. Alien numbers being sequential could make possible estimations of the population of detained children, and this estimation could serve as a means to corroborate or interrogate reported numbers of people in Office of Refugee Resettlement custody. In light of current and widespread reports of abuse of migrants in border agency custody in the United States, third-party validation of government-reported data would be valuable, and there would be enormous public interest value in knowing the total number of children in detention.
At the same time, patterns in alien number assignments pose risks to the security of the data of non-US citizens. Knowledge about alien number assignments could make it possible to make conclusions about detained migrants over time, but it also could have significant implications for the security of immigrants’ data and identities.
This report details the dataset used for this analysis, the investigative steps taken to understand the relationship between alien numbers and dates of entry into detention facilities, possibilities for estimation, and the potential leakage of sensitive data. Read the full report here.
This research was made possible through support from the Ford Foundation. For more information about HRDAG’s supporters, please see our Funding page. Huge thanks are also due to my HRDAG colleagues Valentina Rozo Angel, for her feedback and R-debugging tips, Tarak Shah for his time helping to troubleshoot, and Dr. Patrick Ball for his ongoing guidance and mentorship. Credit is also due to my friends and colleagues at Lucy Parsons Labs, without whom this work would never have seen the light of day.
Photo by Ozzy Trevino.