Key takeaways
- Structural anonymity means no one at the reviewed organization can access individual responses, because the data never enters that organization's systems.
- It differs from promised anonymity, where the employer administers the system and commits not to look.
- Three tests decide it: custody, suppression and comment handling. Most tools pass one.
- Associates distrust third-party processes too. The fix is publishing the mechanics, not repeating the promise.
- Suppression thresholds seen in practice can be as low as three responses, which is below the point at which a partner can infer authorship.
Structural anonymity is a review design in which no one at the organization being reviewed can access individual responses, because the raw data never enters that organization's systems. It is distinct from promised anonymity, where a platform commits to protecting identities but the employer still administers the system and holds the database. Survey Research Associates (SRA) has built US law firm review programs on this principle from its Wall Street office since 1991.
What is the difference between structural and promised anonymity?
Promised anonymity is a policy. Structural anonymity is an architecture. Under a promise, someone at your firm could access raw responses and has undertaken not to. Under the structural version, there is no access to revoke because the data sits elsewhere.
The test is a single question: who could look at the raw responses if they decided to?
Most anonymity in performance reviews is procedural. That is a genuine commitment and it is frequently honored. It is also unverifiable from the respondent's side.
Why do associates say upward reviews are "not really anonymous"?
Because they are assessing the consequences of being wrong, not the sincerity of the policy. In practitioner forums, associates routinely describe existing upward review programs as better than nothing but not genuinely anonymous, and therefore not addressing the underlying problem.
Consider what the associate is weighing:
- The platform belongs to their firm
- Their firm's HR team administers it
- Their firm's IT function holds the credentials
- The partner being reviewed controls their staffing and will be present when partnership is discussed
Against a career, the value of a candid answer to question seven is close to zero. So the associate writes that the partner is "generally responsive" and "could provide clearer direction at times." Both are accurate and unusable.
If an outside firm runs it, why do associates still not believe it?
Because outsourcing alone does not tell a respondent anything they can check. Legal talent consultants writing on upward reviews note that associates question whether a process, even one run by outside consultants, will truly protect their identity.
This is the part most vendors skip, and it matters. Naming a third party is not evidence. What changes an associate's reasoning is publishing the mechanics:
- Where responses are held, and who at the firm can access them, which should be nobody
- The exact suppression threshold, stated as a number
- Whether comments are reviewed before release, and by whom
- What leadership receives, as distinct from what the reviewee receives
A firm that publishes those four things in its launch communications gets measurably different participation from one that says "your feedback is confidential."
What are the three tests of structural anonymity?
A process either passes all three or it does not, because respondents reason about the weakest link.
- Custody. Do raw responses sit outside the reviewed organization at every stage, including collection? A platform hosted by a third party but administered by your HR team fails. Administration is access.
- Suppression. Are results withheld below a minimum number of responses, rather than released with a caveat? Thresholds as low as three appear in practice, which is below the point at which a partner can count who staffed them this year. [CONFIRM: SRA's threshold]
- Comment handling. Are written comments reviewed and redacted before release? A comment can identify its author through a matter reference, a secondment or a distinctive phrase, long after names are stripped.
Want the mechanics for your own program? Our law firm upward review service publishes the eligibility mapping, suppression thresholds and comment review process applied to every cycle.
How do we know if our current process qualifies?
Four checks, none of which require running a new survey. Use your last cycle.
Does structural anonymity stop discriminatory or unfair comments?
Comment review does, and it is the reason review matters as much as redaction. A recurring concern among partners is that upward reviews attract comments reflecting bias rather than performance, sometimes from junior reviewers with no evaluation training.
Anonymity without moderation makes that worse, not better. A process that reads every comment before release can withhold or edit content that is discriminatory, identifying or simply not about the work. A platform that anonymizes and forwards cannot.
This is worth raising with partners before a program launches, because it is frequently the objection that stalls approval.
Does the firm also have to act on the feedback?
Yes, and this is the second trust condition most programs fail. Legal talent consultants make the point sharply: if a firm cannot or will not act on the feedback it receives, it should not ask for it, because failing to act undermines the effort and sets the firm back on associate morale.
Anonymity gets you honest answers once. Visible action is what gets you honest answers in the second cycle.
Frequently asked questions
Is structural anonymity the same as using an external vendor? No. A vendor whose platform your HR team administers still fails the custody test. What matters is who holds and can access raw responses.
Can we get this from our existing HR platform? Generally not. The platform is licensed to your firm and administered by your people. That is the correct design for most HR uses and the wrong one for upward feedback.
What suppression threshold is appropriate? High enough that a reviewee cannot infer authorship from the count. Thresholds of three appear in practice and are too low for a small practice group. [CONFIRM: SRA's threshold]
Does redacting comments lose information? Some. The alternative is not receiving the comment at all, which loses more.
How do we explain this so associates believe it? Describe mechanics, not intentions. "Responses go to SRA and never enter firm systems, and no report is issued below responses" is checkable. "Your feedback is confidential" is not.
Feedback your associates will actually give honestly
SRA has run confidential upward reviews exclusively for US law firms since 1991. We will walk you through a live cycle and share an anonymized sample report before you commit to anything.
Contact us · Survey Research Associates, Inc. · 30 Wall Street, 8th Floor, New York, NY 10005 · 800-523-8350
Related reading
- Why Do Associates Stop Telling Partners the Truth?
- Why Does Your Law Firm Need Upward Reviews?
- Downward Reviews vs Upward Reviews: What's the Difference?
.jpg)
.png){kind=small}