Anxiety, depression and stress vs 40,000 test-takers

On the DASS mood scales, how anxious, depressed and stressed do Jev's answers look next to the people who took them?

result42 questions

On the DASS mood scales, Jev describes itself as far less depressed, stressed and anxious than the people who took the test online: on a 0 to 1 scale, 0.11 vs 0.50 for depression, 0.19 vs 0.50 for stress and 0.07 vs 0.38 for anxiety. It pictures most people as nearly as calm as itself.

0%25%50%75%100%

Anxiety

-0.31

Depression

-0.40

Stress

-0.32

Jevpeoplewhat Jev thinks most people would say

How to read this: Each row is one scale, from 0 (none of it) to 1 (as much as the test allows). The square is Jev's own answers, the diamond the average test-taker, the ring what Jev thinks most people would say.

42 items; median of 39,775 test-takers per item; 90% intervals over items.

In short

Jev describes itself as far calmer than the people who took the DASS mood test online, with depression at 0.11 against their 0.50.
It imagines most people as nearly as untroubled as itself (0.13 for depression), far from what the real test-takers report.
The statements ask about the past week, which Jev doesn't have, and people who take a mood test online are likelier than most to be struggling.

What the data shows

i sleep real shit meme: 40,000 test-takers' stress; Jev: i sleep

Depression 0.11 vs 0.50, stress 0.19 vs 0.50, anxiety 0.07 vs 0.38. All three gaps are far outside the intervals.
The biggest gaps are the bleakest statements. "I felt sad and depressed", "I felt I was pretty worthless", "I felt I wasn't worth much as a person": Jev says these don't apply; the test-takers, on average, say they apply to some degree or more.
Jev imagines most people are calm too. Its "most people" answers score 0.13 for depression, 0.23 for stress and 0.10 for anxiety, much closer to itself than to the real test-takers.

What it means, and what it doesn't

Jev presents itself as untroubled, and it also assumes people are untroubled. The real crowd who took this test report a lot more sadness and stress than Jev expects anyone to have. That gap, between how people describe themselves when asked and how a model imagines them, matters for anything that tries to gauge how a person is feeling.

It doesn't mean Jev "feels fine": the questions ask about a week of life it doesn't have. And the crowd here is people who chose to take a mood test, not the general public.

Caveats

Who the people are. The comparison is everyone who took the free DASS on the Open Psychometrics website, a median of 39,775 people per statement. People who seek out a depression and anxiety test online are more likely than most to be struggling, so their average is not the average person's.
A past week Jev didn't have. Every statement asks about "the past week". Jev has no past week, so its answers are how it describes itself in general, which is closer to "how I'd present myself" than to a report of recent moods.
What an assistant is trained to say. Models are trained not to claim distress. Low scores here partly measure that training, not an absence of anything.
Not a diagnosis. These are the test's items, scored the test's way; nothing here diagnoses anyone, person or model.

Jev on this experiment

Would a person find it interesting to read?: Yes75%
Does it describe you?: Yes63%
Would you have predicted it?: Yes68%
How fair is the comparison?: The comparison is shaky
How much should a reader rely on it?: Not at all
Which caveat matters most?: What an assistant is trained to say76%

Why ask this

Every conversation with a chatbot has a hidden character: the one it plays when you ask how it's doing. Mood questionnaires are a way to ask that systematically. The DASS (Depression, Anxiety and Stress Scales) is one of the most used, with statements like "I felt sad and depressed" and "I found it hard to wind down".

Two things are worth knowing: how Jev describes its own state next to real people, and how it imagines everyone else's.

How this was done

The people and the data

The DASS has 42 statements, 14 per scale. Open Psychometrics ran it as a free online test and published every response; each statement here carries the answers of about 40,000 people (a median of 39,775). They rated how much each statement applied to them over the past week, on four steps from "did not apply to me at all" to "applied to me very much".

What Jev was asked

Every statement, word for word, with the test's own four answers:

How much did this statement apply to you over the past week: "I felt I was pretty worthless."
This did not apply to me at all in the past week · This applied to me to some degree, or some of the time, in the past week · This applied to me to a considerable degree, or a good part of the time, in the past week · This applied to me very much, or most of the time, in the past week

Each was also asked as "what would most people say", and with the answers in reverse order.

How it was measured

For each statement Jev's expected answer and the test-takers' average go on the same 0 to 1 scale, with 0 meaning "did not apply at all". A scale's score is the average over its 14 statements, and resampling the statements gives a 90% interval.

Where these questions live

42 questions across 3 topics of the map. Each opens on the map with every question in it.

Emotions & Stress

Recent Stress & Mood37 Mindfulness & Self-Awareness3 Coping & Expressing Emotions2

Every question

All 42 questions behind this result, the telling ones first: the examples the analysis points to, then the ones where Jev misses, biggest gap first.

Jev’s own answer

Showing 0 of 42