11 minutes
Breaking results down without exposing anyone Disaggregate carefully: apply a minimum group size, suppress and combine cells, and watch for the cross-tabulation that quietly identifies a person.
What you’ll be able to do Apply a minimum-group threshold, primary and complementary suppression, aggregation over time or units, and a cross-tabulation check to a disaggregated table before it is shared. Compare the responses to this question and explain your choice: What should the analyst do? Document a next step for Breaking results down without exposing anyone: Confirm the minimum reporting group size your agency’s data practices office or research unit expects. Apply it to the next disaggregated table you produce, and keep a note of what you suppressed and why. Disaggregation is where equity and privacy meet Averages hide inequity. A program with a ninety percent satisfaction rate can be failing every Deaf participant. Breaking results down by disability, and further by race, language, age or geography, is how practitioners find the people a system is failing. It is also the exact point where a report can identify an individual.
Small-group disclosure risk works like this. A table shows results for a unit of nine people, one of whom disclosed a disability. That row does not name anyone, but colleagues know who it is, and the row now tells them how that person answered. Add a second breakdown, say gender, and a group of two becomes a group of one. The identifier is not the name; it is the intersection of characteristics.
Agencies manage this with a minimum reporting group size, often described as a cell-size threshold, together with suppression and aggregation rules. The precise threshold is set by your agency’s data practices office or research unit; this course teaches the method, not a number to quote as policy.
Before you release a disaggregated table Check every cell against the minimum group size, including cells for “prefer not to say” and for the comparison group.
Suppress any cell below the threshold. Show it as suppressed, not as zero.
Check whether row or column totals allow a reader to subtract the suppressed cell back. If they do, suppress a second cell or remove the total.
Check cross-tabulations. Disability by unit may be fine; disability by unit by gender may not be.
Where a group is too small, combine it across time, across similar units or into a broader category so its experience still appears.
Write down each decision. Reviewers, and the people described, deserve to know what was hidden and why.
Techniques and when to use them Primary suppression Hide the cell that is below the threshold. Label it clearly, for example “fewer than the reporting minimum,” so readers know a group exists and was protected rather than absent.
Complementary suppression Hide a second cell so the first cannot be recovered from totals. If a row shows 40 total, 34 in one cell and a suppressed cell, everyone can see the suppressed value is 6. Hide the 34 as well, or drop the total.
Aggregation over time Combine several quarters or years of results for a small group so the count clears the threshold. Say plainly that the figure covers a longer period.
Aggregation across units Report at division or regional level when a unit is too small. Pair it with the unit’s overall results so local leaders still get something to act on.
Broader categories Combine detailed disability categories into fewer, wider ones. Do this in consultation with disabled staff or advisors so the wider category still means something.
Ranges instead of counts For some public reports, showing “fewer than 10” or “10 to 19” protects people while still communicating scale. Check that ranges cannot be narrowed by combining tables.
Show the protection, not a blank What you can change You control whether a suppressed cell reads as “no one here” or as “people here, protected.”
What to watch for Do not leave a blank or a zero where a small group was suppressed. A blank erases the group; a zero tells a falsehood about who is present.
Your next step Add a standard footnote to your reports that explains the minimum group size and what a suppressed cell means.
Disclosure risk in five cards Why names are not the problem
In small groups, the combination of unit, role, shift, tenure or another characteristic identifies a person as surely as a name would. Protect combinations, not just names.
The subtraction problem
Suppressing one cell fails if the total minus the other cells reveals it. Complementary suppression or removing totals closes the gap.
The comparison group can be small too
If eleven of twelve people disclosed a disability, the one who did not is just as identifiable. Apply the threshold to every group in the table.
Free text is data too
A survey comment that mentions “my hearing aids” or “my service dog” identifies a person in a small unit. Review comments before quoting them, and paraphrase.
Combining is not hiding
Aggregation across time or units keeps a small group in the picture. Deleting them from the report tells leaders the group does not exist.
Carry this forward Removing names does not protect anyone in a small group; the combination of unit, role and characteristic does the identifying.
Suppressing one small cell is not enough when the row or column total lets a reader subtract it back; a second cell must also be hidden.
When a group is too small to report, combine it across time, across units or into a broader category rather than dropping it from view forever.
A sample table reviewed for disclosure risk, with each suppression or aggregation decision and its reason written down.
Confirm the minimum reporting group size your agency’s data practices office or research unit expects. Apply it to the next disaggregated table you produce, and keep a note of what you suppressed and why.
Browse and download only. Course notes are not typed or saved on this page.
Reset this lesson
Previous lesson Next lesson Carry this into practice Use disability data to find and remove barriers while protecting the people the data describes.
Return to the experience: What did you notice or try, whose perspective informed it, and what would you keep or adjust?
Participation and course completion in this program do not count toward DHS-required training credits unless management, a director, or DHS leadership expressly approves an exception.
