Counting victims in an archive requires more than counting photographs

The leaked archive contained more than 33,000 photographs of detainees’ bodies. That was not the number of people represented: several images could show one person. ICIJ’s team used the folder structure and detainee identifiers to establish a count exceeding 10,200 bodies.

The reporting explains an unusually consequential data-cleaning process. Arabic directory names contained dates and one or more detainee numbers. Translating, separating, and manually counting those identifiers made it possible to distinguish records from individuals.

A separate random sample of images was examined by trained journalists to investigate what the photographs showed. That analysis answered a different question from the enumeration of people; neither should be casually substituted for the other.

The account also documents the psychological burden of reviewing such material. Its methodological value is precision under difficult conditions: explain how a number was produced, preserve the distinction between observation and inference, and avoid allowing the scale of a leak to obscure individual lives.