Why it matters
The connection may rely on genetic comparisons, metadata, or other information. A credible account explains what was available and how strong the resulting association was, rather than suggesting that every unnamed record is equally identifiable.
A practical example
Imagine a public dataset uses study numbers while a second source contains overlapping identifying details. Linking the two may narrow a record to a person, even though neither source alone displays the full connection.
An important distinction
Identifying a person differs from inferring membership in a dataset or predicting a trait. These are distinct findings with different evidence requirements. Research demonstrations should be described with their actual assumptions and limits.
Sources and further reading
- PRIMARY SOURCE NIST SP 800-188: De-identification
The example above is hypothetical. Related cases provide context; they do not imply that every case involved this mechanism.