Identifiers to Consider in Human Subjects Research (Non-HIPAA)

This document lists categories of identifiers that may make data identifiable in human subjects research projects. Researchers should consider whether any of the following identifiers will be collected or generated as part of their study. This is not a comprehensive, exhaustive list – rather, it is a list of examples of identifying information for your consideration as you build out your human subjects research project.

A. Direct personal identifiers (non-HIPAA–specific)
  • Full Names
  • Personal usernames or handles tied to real identity 
  • Student ID numbers
  • Employee ID numbers
  • Organizational ID numbers (e.g., union, professional association)
  • Utility account numbers
  • Library or campus card numbers
  • Research participant registry numbers
  • Court case numbers tied to individuals
  • Immigration file or case numbers
  • Prisoner numbers or correctional identifiers
  • Any study-generated code or number that can be linked back to the individual through a key
B. Visual and audiovisual identifiers
  • Full-face photographs
  • Partial-face images that enable recognition
  • Video recordings
  • Audio recordings containing an unmodified voice
  • Images showing distinctive tattoos, unusual scars or marks, or distinctive clothing
  • Screenshots or recordings of virtual meetings or online platforms
C. Geolocation and movement data
  • GPS coordinates
  • Detailed movement traces
  • Heat maps of movement that identify a household or workplace
  • Ride-share, transit, or travel histories
  • Location pings from phones, apps, or wearables
  • Small-area geographic identifiers that allow singling out a person or household
D. Online and digital identifiers
  • Device identifiers
  • Browser fingerprints
  • Advertising IDs
  • Metadata embedded in images or files
  • Chat logs
  • Forum or gaming platform handles
  • Platform-specific participant IDs
  • Clickstream or activity logs tied to the same user
E. Genetic and biospecimen information
  • Whole genome or exome sequence
  • Other uniquely identifying genetic markers
  • Family pedigrees or rare variant patterns
  • Biospecimens linked by a key to an individual
  • Familial relationships enabling identification of relatives
F. Contextual or quasi-identifiers (become identifying in combination)
  • Exact occupation or job title
  • Department, lab, or workgroup
  • Military rank and assignment
  • Membership in small organizations or communities
  • Rare disease or condition in a small geographic area
  • Highly specific demographic combinations (for example, age plus small location plus occupation)
  • Household structure
  • Specific educational program or classroom
  • Detailed employment history
  • Unique life events timeline
G. Behavioral and usage pattern identifiers
  • Typing cadence or keystroke dynamics
  • Gait or movement pattern data
  • Voice or speech pattern data (even without content)
  • App use logs tied to a persistent identifier
  • Distinctive online activity patterns
H. Narrative content that can inadvertently identify a person
  • Participant narratives describing unique events
  • Small-community references
  • Descriptions of lawsuits, news coverage, or public incidents
  • Quotations that can be searched online
  • Occupational anecdotes with identifiable detail
I. Research operations identifiers
  • Randomization lists
  • Re-identification or linkage keys
  • Codes used across multiple datasets
  • Master subject lists
  • Crosswalk files