What are key considerations for de-identifying data in DEOCS reports?

• A. Remove unit names when sample sizes are small.
• B. Remove unit names when sample sizes are small, mask or redact direct identifiers, and ensure aggregate categories do not allow deduction.
• C. Keep unit names visible to provide context.
• D. Publish raw data tables.

Answer: (B)

Protecting respondent confidentiality in DEOCS reports relies on de-identifying data so individuals cannot be traced. When a unit’s sample is small, listing the unit name can make it possible to connect a respondent’s answers to that person, so unit identifiers are removed in those cases. In addition, any direct identifiers—names, contact details, employee IDs—should be masked or redacted to prevent obvious linkage. Finally, even when data are aggregated, care must be taken that the combination of categories does not enable someone to deduce who answered. This often means suppressing or grouping small cell counts and avoiding publishing tables that could reveal a single individual’s responses. Together, these steps protect privacy while still allowing useful reporting of overall trends and patterns.