Clinicians

Collateral Informants: How to Get Useful Teacher and Partner Ratings

September 12, 2026 · 6 min read

Last updated September 12, 2026

An ADHD diagnosis built on one person’s self-report is a hypothesis, not a finding. The criteria require symptoms in two or more settings, and the person being evaluated is a poor witness to their own behavior in at least one of them. Collateral informants supply the missing view. Done carelessly, they add noise: a teacher who has known the child three weeks, a partner coached on what to say. Done well, they are the strongest evidence in the file.

Why self-report is insufficient

Adults with ADHD tend to under-report symptoms relative to observers, and adults seeking a diagnosis may over-report, sometimes without intending to. Children under about eleven are unreliable self-reporters on attention altogether. Sibley and colleagues found that in young adults, informant reports and impairment ratings identified ADHD more accurately than self-report on the same items. Barkley’s adult work reaches the same conclusion: the partner or parent rating is closer to observed behavior than the client’s. Collateral is not a nice-to-have. It is where the diagnosis lives.

Choosing informants for a child

The classroom teacher who has had the child for at least six to eight weeks, and if the child has multiple teachers, two of them from different subjects. A parent, and both parents separately when possible, since discrepancies between them are informative. For adolescents, add the teen’s own rating and expect it to be lower than everyone else’s. Avoid informants whose relationship to the child is primarily about the referral (a tutor hired for the problem) or who have a stake in a particular outcome.

Choosing informants for an adult

A partner or cohabitant who sees daily functioning. A parent, for childhood symptoms specifically; the adult ADHD scales include a retrospective childhood form for this purpose. Where neither is available, a sibling, close friend, or longtime coworker. Ask the client to name someone who will be honest rather than kind, and say why: the evaluation is only useful if it is accurate.

Instructing the informant

Send the scale with a short cover note, and give the same note to everyone. Rate what you have actually observed in the last six months, not what you have been told. Rate the person compared with others the same age, not compared with how they used to be or how you wish they were. If you have not observed an item, leave it blank rather than guessing. Do not discuss your ratings with the person being evaluated until after they are returned. That last instruction matters more than it seems; ratings completed together at the kitchen table converge on the client’s own view and lose their value.

Reading disagreement

Informants will disagree, and the temptation is to average. Resist it. Parent-high, teacher-normal in a child suggests a home-specific problem, a parent-child relationship issue, or a parent who is themselves anxious or has ADHD and is rating against an atypical baseline; it is not consistent with pervasive ADHD until you understand why. Teacher-high, parent-normal often means the demands of school expose a deficit the home does not, which is consistent with ADHD or with a learning disorder. Partner-high, self-normal in an adult is the classic ADHD pattern. Self-high, partner-normal should send you back to the differential, particularly anxiety and depression.

Documenting it

Name each informant, their relationship, how long and in what settings they have known the client, and the instrument and version they completed. Report their scores separately, never combined, and write a sentence about agreement or disagreement and what you make of it. A report that does this is nearly unassailable; a report that says ‘collateral information supports the diagnosis’ without any of it is not. The report-reading guide describes what your readers will be looking for.

The bottom line

Two settings, two informants minimum, chosen for observation time and honesty, instructed identically, rated independently, and reported separately with their disagreements interpreted rather than averaged. That is the evidentiary core of an ADHD evaluation.