Why do people answer mental health surveys to look better than they feel?
Many people shade their answers on mental health surveys so they look more okay than they feel. In one Army study, soldiers reported depression, PTSD, and suicidal thoughts 2 to 4 times more often on an anonymous survey than on their official, named screen. People tend to share more when they feel no one is judging them.
What does this mean for you?
The same pattern shows up far from the Army. In the Veterans Health Administration, Samantha Louzon and colleagues followed 391,492 patients who took the PHQ-9. Higher answers on its suicide question did predict higher risk. But 71.6% of the 310 people who died by suicide had answered "not at all" on their most recent PHQ-9. A study of 76 inpatient suicides by Katie Busch and colleagues found 78% denied suicidal thoughts at their last check. It even happens inside therapy. Matt Blanchard and Barry Farber found 93% of 547 clients had lied to their therapist at some point, often by pretending therapy was working.
Does the listener matter? It seems to. In a 1996 meta-analysis of 39 studies, Suzanne Weisband and Sara Kiesler found people disclosed more on computer forms, most of all compared with face-to-face interviews. At USC, Gale Lucas and colleagues had 239 adults talk to a virtual human interviewer. Those who believed it was fully automated reported less fear of opening up and less impression management. What mattered was the belief, not who was really running it. A later small study of 24 National Guard members found they reported about 3 times more PTSD symptoms to the virtual interviewer (0.79 on average) than on the official screen (0.25).
Is the evidence strong? It is strong that people hold back on named, high-stakes screens. It is more mixed on how much a computer helps by itself. In the Warner study, the two groups were not exactly the same people. The first Lucas military study was small, and in a larger sample of 126 the overall gain fell just short of statistical significance. The computer effect has also gotten smaller in newer studies, though a 2015 meta-analysis by Gnambs and Kaspar of 125,672 people still found more disclosure on computers than on paper.
Surveys still matter. In the same VA study, higher scores did predict higher risk. Validated tools like the PHQ-9 are reliable and useful. They just work best when they are not the single thing a clinician sees.
A survey is a snapshot taken in front of an audience. When people feel watched, judged, or worried about what happens next, the snapshot tilts toward "I'm fine." That is human. Most of us want to look like we are coping. The research points to what helps people open up: privacy, not feeling graded, and a conversation that builds trust. For a person in the gap between sessions, that matters. The hardest days often happen far from the therapy room. If a waiting room form is the only record of those days, a lot gets lost.
If you are struggling: if you are having thoughts of suicide or self-harm, you do not have to carry it alone. In the US, call or text 988, the Suicide and Crisis Lifeline, or go to your nearest emergency room. Outside the US, contact your local emergency number.
The takeaway
- If a survey score feels "too good," it may be worth a second, gentler conversation.
- Privacy and trust change what people share. Tell people who will see their answers, and why.
- Use validated scores and ongoing, honest conversation together. Each fills gaps the other leaves.
This is the work Asendar™ was built for: pairing validated assessments and the Mental Readiness Score with consented, private, between-session conversations, so a provider sees more than a waiting room snapshot. We are mapping how people are doing mentally and emotionally in real time through the conversations they are having with AI.
Take your Mental Readiness Score