Are Online ADHD Tests Reliable? What a Score Can Tell You

Before booking my appointment I filled in those tests too. Several of them. All came out high, and that evening I was stuck in a strange place: I was relieved, thinking "there's the proof," and I was also angry with myself, thinking "this doesn't mean anything."

Both turned out to be true. They do not diagnose, but they are not useless either. Without saying both sentences together, the subject cannot be understood.

What they were designed for

Most of these tools were made for screening. Screening and diagnosis are different jobs, and the reason they get confused is that both are made up of the same questions.

The best known is the ASRS. This scale, developed by Kessler and colleagues (2005), was designed for the World Health Organization for population-wide screening. Its purpose is this: to flag, within a large population, the people for whom a more detailed evaluation would be worthwhile.

So the scale's job is to point to a door. Going through the door is a separate job. In the glossary this is called a screening tool.

A screening tool says "there is something here worth a look." It does not say "this something is ADHD."

How wrong even a good screening tool can be

When I first read this part, the picture in my head changed, so I want to write it with numbers.

The six-question version of the ASRS updated for DSM-5 performed as follows in a population sample: it catches about 91 percent of people who have ADHD and correctly screens out about 96 percent of those who do not. These are very good numbers for a screening tool. But the third number in the same study says more: about 67 percent of those whose test was positive actually have ADHD (Ustun et al., 2017).

So even with a well-designed, scientifically validated scale, in one out of every three people who test positive the diagnosis is not confirmed. In the glossary this is called a false positive, and its cause should be sought not in the test being bad but in the mathematics: the rarer the condition being looked for in the population, the larger the share of positive results that turn out to be false.

Most of the tests circulating on the internet do not even have a validation study of this kind.

Why they seem to come out positive for everyone

There are a few real reasons for this impression.

The symptoms overlap. Distractibility, forgetfulness and restlessness are not seen only in ADHD. Anxiety disorders, depression, sleep disorders, thyroid problems and iron deficiency can produce similar pictures. The review by Katzman and colleagues (2017) shows how common this overlap is in the clinic. I covered the distinction from anxiety in the article on ADHD or anxiety.

The questions are broad. To the question "Do you often lose your things?" almost anyone going through a tiring period can say yes.

The interpreter is biased. A person looking for the test is already looking out of a suspicion. They read their own answers in the direction of that suspicion. I saw this clearly in myself: on the questions where I was caught between "sometimes" and "often," I always rounded to the same side.

Some tests are commercial. If a service is being sold at the end of the result, the test may tend to give a positive result. Who published it and what stands behind it are worth looking at.

What happens when it comes out low

This side is talked about less, and I think it is more important.

Screening tools work better in people whose symptoms show outwardly. In people who have learned over years to suppress their symptoms, who do not cause trouble from the outside and whose inattentive presentation dominates, they can miss it. I wrote about this picture in detail in the article on inattentive ADHD.

The practical counterpart is this: a low test result does not rule out ADHD. If there is a real struggle in your life, do not give up on applying because of a low test result.

How diagnosis is made

In the process described by the European consensus statement (Kooij et al., 2019), a scale alone is not enough. The evaluation includes the following:

• Developmental history. The symptoms are required to have been present in childhood too.
• The two-settings rule. The symptoms must be seen in at least two areas, not just one.
• Functional impairment. The symptoms must be affecting life in a measurable way.
• Differential diagnosis. Whether there is another condition that explains the picture better.
• If possible, another source. The observation of family or someone close.

No online test can do any of these steps. The decision about diagnosis and medication is the job of a psychiatrist; the details of the process are in the article on how adult ADHD is diagnosed, and where to apply is in the article on where to get an adult ADHD diagnosis.

Which test is better

I cannot give a definitive ranking to this question, but there are criteria to look at, and they can be seen on a test's page in a few minutes.

Who published it? A university, a professional body, a hospital; or a site whose name nobody knows.

Which scale is it based on? A scale with a name and a source like the ASRS, or a list of questions that does not say where it came from.

How does it describe the result? If it gives a number like "you are 87 percent ADHD," stay away; no screening tool can say such a thing. A good test result says "this score may be grounds for an evaluation."

What happens at the end? If it directs you to a product, a course or an "online diagnosis" service, the job of that test may be to convince you.

What does it do with your data? Looking at what the site collects before entering your answers is a reasonable habit, because it is information that can also go somewhere unrelated to ADHD.

So what should you do with the result

The useful way to use it is this: take the result not as an answer but as a note.

When you go to the appointment, take these with you as well:

• Which items you gave high scores to on the test
• What you remember from your childhood, and report cards and teacher comments if you have them
• Which concrete things are falling short: missed deadlines, finished tasks, unpaid bills
• An observation from someone in your family

This list makes a fifteen-minute interview far more productive. Giving lived examples instead of counting off a list of symptoms makes the evaluation easier.

The one question to ask while filling in the test. As you answer each item, do not ask "does this fit me?"; ask "what did I lose in the last year because of this?" Mark the items for which you have an answer and write that example next to it. What the physician needs is not the score. It is those examples.

Social media diagnosis

There is one more heading: diagnosing yourself from short videos.

There is a useful side to this content; many people say it is where they recognized themselves for the first time. I am one of those people and I do not want to belittle it. But it has a trap too: generic descriptions fit almost everyone. This is what appears in the glossary as the Barnum effect.

The distinction is here: if a description makes you say "I'm like that sometimes too," that information is weak. If it makes you say "this is why I lost that job," the information is strong. I discussed the logic of the medium itself in the article on why ADHD has become so widespread on social media.

The sentence I use

I did not give up on tests altogether. I only changed what I ask them.

I used to ask "do I have ADHD?" and expect an answer from a screen. Now I ask this: is this result a sufficient reason for me to go and see someone? A screening tool can really answer that question. It never could have answered the other.