Public Access Test for a Service Dog: The Certificates Are Worthless, the Test Isn’t

Updated September 20, 2026 · first published September 7, 2026Rewritten in full against the current regulations

Somebody asks me about the public access test for a service dog most weeks, and it is nearly always the same question wearing a different coat: do I have to have this, and where do I buy it?

The answer to the second half is that you cannot buy it, and the people offering to sell it to you are selling you nothing. The answer to the first half takes a little longer, and it is not the answer either side of this argument usually gives.

I have been a handler for twelve years. Mr. Big never sat a formal public access test in his life, because by the time I needed one it was not publicly available anymore. So we improvised our own. Here is what is actually in the test, what I did instead, and what the certificate industry is charging for a piece of paper that does nothing.

The short answer: is access testing mandatory?

No. There is no public access test requirement anywhere in United States law. Not federal, not state, not municipal. No business can ask you to produce a test result, and no test result gives you a right you did not already have.

I want to be careful here, because this is the part where most articles stop and where most of them get it wrong in the other direction. The test is not a scam. The test is a good piece of work. What has been built on top of it is the problem, and the two are not the same thing.

So here is the position I have landed on after twelve years of doing this:

Nobody can make you take a public access test. You should probably take one anyway.

The rest of this is why, what is actually in the test, what I did with Mr. Big instead, and what the people selling certificates are charging for a piece of paper that does nothing.

What the public access test actually is

The document everyone means when they say “public access test” is the Assistance Dogs International Public Access Certification Evaluation. The version that circulates is stamped ADI.PATMAY2014. It has fourteen sections.

I am not going to host the PDF. It is ADI’s copyrighted document and they pulled it from public view deliberately. But I am going to describe what is in it, because the items themselves are not a secret and should never have been treated as one. This is basic obedience in public. It is not proprietary knowledge. If you are training your own dog, you need to know what the bar looks like.

The fourteen sections cover:

• Controlled unload from a vehicle. The dog waits, comes out on command, stays with you.
• Crossing a parking lot. Heel position, no pulling, no bolting at traffic.
• Entering a building. Through the door under control, no charging ahead.
• Heeling through a building. Aisles, corners, changes of direction.
• Heeling through a crowd. People moving around you in a confined space.
• A six-foot recall on lead. The dog comes when called with distractions present.
• Three sits. On command, held, including with food nearby.
• Three downs. Same, including a down near a dropped piece of food the dog must ignore.
• Noise distraction. Something loud and sudden. The dog may startle. It may not panic, bolt or bark.
• Restaurant behavior. Under the table, settled, for the length of a meal.
• Recovering a dropped lead. The lead hits the floor and the dog does not leave.
• Being left with someone else. You hand the lead to a stranger and walk out of sight.
• A controlled exit. Out of the building the same way you came in.
• Overall team relationship. A judgment call on how the two of you work together.

The scoring works like this. Some items are scaled and have to come back as Always or Most of the time. The binary items need 80% yes. A handful of items are marked compulsory and every one of those has to be a yes regardless of the total.

I will tell you what I think of that scoring in a minute, because I do not entirely agree with it.

One thing about the document that almost nobody mentions: it does not say who is allowed to administer it. Nowhere in the fourteen sections is there a line saying this must be run by a certified evaluator, a trainer, a member organization, anybody. That silence is doing a lot of work in this article and we will come back to it.

What is printed on the test itself

Two lines are printed on the ADI test document. Here they are:

“Passing This Test Does Not Imply Certification By Assistance Dogs International”

“Does Not Assume Responsibility For Use Of This Test”

The people who wrote the test printed on the test that passing it certifies nothing. That is about as clear as an organization can be.

ADI’s current public pages say the same thing in longer form:

“ADI does not directly provide, train, certify or register individual assistance dogs.”

The test is no longer publicly available. Cooperative Paws, run by a credentialed trainer, reports that ADI withdrew it from public availability, and ADI’s own website is consistent with that — their full standards are now described as available only to accredited members and candidate organizations, and their public pages point to no downloadable test. The copies floating around online are third-party mirrors of the 2014 file.

I will be honest that this bothers me, and it is the one place where I part company with ADI.

They want service dogs performing to this standard, but they will only hand the standard to their own list of training providers. Federal law says my handler-trained dog has exactly the same rights as one from a program. So why is the yardstick behind a door?

I do not think that is a conspiracy. I think it is an organization of training programs behaving like an organization of training programs. But the effect is that the people who most need to know what good looks like — owner-trainers, working alone, usually broke — are the ones who cannot get it. That is a little weird to me, and it opens a doorway straight to paid training.

What the law actually requires, which is two things

Set the test aside for a second. Here is the entire legal bar for a service dog in a public place, from the Department of Justice:

“If a particular service animal is out of control and the handler does not take effective action to control it, or if it is not housebroken, that animal may be excluded.”

Out of control, and not housebroken. Two conditions, judged on the day, by the person in front of you.

Not fourteen sections. Not 80%. Not a card, not a vest, not a registry entry.

DOJ is equally direct about paperwork:

“Covered entities may not require documentation, such as proof that the animal has been certified, trained, or licensed as a service animal, as a condition for entry.”

“These documents do not convey any rights under the ADA.”

And on whether you are allowed to train the dog yourself:

“No. People with disabilities have the right to train the dog themselves.”

No state requires a public access test either. Access under state law rests on the dog’s training status, not on a document — you can check that state by state in the Animal Legal and Historical Center’s table. I have written up the ones that matter most on my ADA service dog laws page.

The four states that mention identification, and why none of them is a test

People send me these as if they are counter-examples. They are not, but they are worth knowing about because they are where the confusion starts.

Connecticut refers to a harness or an orange lead. California has a county tag scheme, and the part people quote at me actually applies to trainers — I have covered it in detail on the California service dog laws page. Idaho and Kentucky provide identification cards for non-disabled trainers, which is a different situation from yours entirely.

Georgia is the one that sits worst against federal law, because its language leans on the dog having been trained by a school. I am not going to interpret that for you and I am not going to tell you how a Georgia court would come out on it. I will tell you the mismatch exists between what that statute says and what the ADA says, and that DOJ’s position on training your own dog is the one quoted above.

And one that matters if you are still training: Hawaii is the only state with no public access protection for a service dog in training. Everywhere else gives you something. If you are working a dog in public before it is finished, read can service dogs in training go anywhere first, because the rules for a team mid-training are genuinely different from the rules for a finished one.

What the certificate sellers are charging

Now the part I have been going back and forth about, because naming names will cost me something.

Search “public access test” and the first page is not ADI, not DOJ, not a trainer. It is registration companies. So let me tell you what they are selling and what it costs, as of September 2026, from their own live pages.

Service Dog Certifications sells something called a Handler-Certified Public Access Test Certificate. It is $39. You tick a box confirming your dog meets thirty-two requirements. Nobody watches your dog. Nobody comes out. You certify yourself, and they print it.

Read that again, because I had to.

It is a certificate for a test you gave yourself, sold by a company that was not there.

On the same site, in their own words: “Service Dog Certifications is not affiliated with any government agency and its products and services do not confer any legal rights on any person.” They also tell you, correctly, that under the ADA staff are not allowed to ask for documentation for service dogs. They are selling you a document while explaining on the same page that nobody can ask you for one.

US Service Animal Registrar runs five packages:

Their own FAQ says: “Does registering give me legal rights under the ADA or FHA? No. Registration is documentation, not a legal status.”

I want you to sit with the shape of that price list for a second, because the shape is the tell. The difference between $99 and $299 is not more rights. It is gold embossing, a lanyard and a luggage tag. Those are the upsells of a novelty gift company, not a credentialing body. Nothing in the top tier makes your dog better trained or your access more secure than the bottom tier, and the bottom tier does nothing either.

NSAR and US Service Animals sell comparable registration kits and certificates. USA Service Dog Registration and Pettable occupy the same page-one real estate.

Here is the move I keep coming back to on this site, and it applies perfectly here: ask who benefits from you believing it. Every one of these companies makes money the moment you become uncertain about whether your rights are real. They do not need you to be wrong for long. They need you to be scared in the doorway of a grocery store, on your phone, for about four minutes.

I have made the certificates argument at length in three other places and I am not going to do it a fourth time here. If that is what you came for, read how to identify a fake service dog and how do I qualify my dog as a service dog. This article is about the test, which is a different and better thing.

What an honest paid test looks like

Because I do not want to leave you thinking every dollar in this space is a con. It is not. Some organizations will genuinely put an evaluator in front of your dog, and the price reflects a person’s time.

Atlas Assistance Dogs works specifically with owner-trainers. Their public access test is $125, with recertification at $75. The full program — intake, assessment, nine months of coached training, then the test — comes to $900, and they offer financial aid. They say plainly on their own site that certification is not required by the ADA and that what they are confirming is the team’s readiness, not a legal status.

Medical Mutts runs an in-person PAT in Indianapolis. Assessment is $120, the first test is $300, and recertification is $120 a year. You get a vest, an ID, a letter, and monthly follow-up with their trainers for the first year.

Service Dog Training School International sells an online, self-administered PAT at $199, reduced from $399, where you submit video of your dog and tutors evaluate it. They state that it is not the same as service dog certification. That one sits in the middle for me — a real human watches real footage, which is worth something, but you are still filming it yourself.

Notice the difference. Atlas at $125 is buying you an evaluator’s attention. Service Dog Certifications at $39 is buying you a printout of your own opinion. The cheaper one is the worse deal, and it is not close.

How I decided Mr. Big was ready

Nobody tested us. I want that on the record before I tell you anything else, because I am not standing here with credentials.

The actual test is not publicly available anymore, so we improvised our own version. And what I did was not a test day. It was several months.

I worked Mr. Big in his vest, on leash, and I took him everywhere I could think of. On leash was not a stylistic choice: the ADA requires a leash, harness or tether unless it stops the dog working. Stores. Shops. City streets. Malls. Parks. Hotels. Restaurants. Outdoor festivals. Farmers markets. Dog parks. In and out of the car, which he loves. Up and down steps anywhere there were steps. Elevators. Escalators. On and off buses. I took him to petting zoo attractions, on purpose, because I wanted animals he could not predict.

Two of those settings deserve a note, because they are where I learned the most. Airports and buses are where public access gets tested hardest — crowds you cannot leave, noise you cannot control, and staff who may have opinions about you. Everything I know about that half of it is in flying with a large service dog and taking a service dog on a bus or train, and I would not have been ready for any of it without the months on the ground first.

I was chasing one specific thing:

Anything that moved, made noise, or had people in it. I did not want him to spook if we hit something new.

The point was never to pass anything. The point was to build enough confidence in him that he could hold his composure even when mine was wobbly. That is the real job. A service dog is not there for your good days.

I decided he was ready for full-time work when all of those conditions were met. Not on a date. Not at a score.

Was there a moment? No. It was gradual. Some things never bothered him for a second. Others took a few tries. My biggest concern the whole way through was that he not react to other animals, to people, and especially to children.

The escalator

We failed at exactly one thing, and it was escalators. I know how that sounds. But I have bad knees, and stairs are not always a choice, so this was not optional for us.

He refused. Several times.

So one day I said to him, alright, we are going to sit down right here. There was a seating area nearby. And we sat and we watched the escalator for about half an hour. Then I took him over and we stood at the threshold, not getting on, just watching it move. Then we left. And I did that several times over several visits.

I was not surprised it took him a while, to be fair to him.

The first time I was on an escalator as a kid it scared me too. It looks like it is going to eat you.

We got there on lots and lots of attempts and lots of praise. That is the whole method. There was no shortcut and there was no certificate involved at any point.

The one section he would have struggled with

Reading back through the fourteen sections now, Mr. Big could do all of it. Genuinely all of it, no hedging.

Except being left with someone else. He did not like that one.

He never acted up about it. He did not whine or break. But here is the thing people who have not worked a dog do not always understand: a service dog knows he is supposed to be with you. He would watch me the entire time I was in eyesight. And the moment I went out of sight, he would sit down exactly where he was and stay there until I came back. Not distressed. Just not willing to be anywhere else.

I am not sure that is a failure. I am fairly sure it is the dog doing his actual job. Which brings me to the parts of the test I would change.

What I would change about the test

All fourteen items are worth doing, and I want to say clearly that for handlers with sight or hearing impairments this list is excellent. No argument from me there at all.

But I would push on four things.

The six-foot recall is too short. Six feet is nothing. I would want to see a longer recall, and I would want an off-leash element in there. Off-leash control belongs in a working dog’s training whether or not it belongs in a test.

Heeling should be tested as a baseline, not just through a crowd. Crowd work is the hard version. If you only test the hard version you learn less than you think about the easy one.

Sits and downs are the most important items on the list and they are scored like everything else. Those are the commands that matter in a place where the dog is in danger, or where you and the dog get separated. They should carry more weight than restaurant manners.

The controlled exit deserves more credit than it gets. Mr. Big’s controlled exit was superb, and it was not obedience. He would get me out of a building if my anxiety climbed past the point where I could bring it down in a few minutes, or if he read a real danger to either of us. The test measures whether the dog can leave a building tidily. It does not measure whether the dog will decide, on his own, that you need to leave.

On the scoring: 80% feels low to me for the non-compulsory items. I would put the compulsory ones at 100%, because that is what compulsory means, and the rest somewhere around 90 to 95%. But I would keep some room in there, because everyone can have an off day, dogs included.

Should you do a public access test?

Somebody owner-training their first dog asks me this and my answer is short.

Absolutely. That is data that helps you fill the gaps before it is a serious matter.

That is the whole case. Not legitimacy, not paperwork, not proving anything to a store manager. Information about your own dog, gathered on a quiet Tuesday, instead of learning it in a Walmart when something goes wrong.

You are not taking the test to pass it. You are taking it to find the thing you did not know was broken.

It is also the clearest line I know of between a service dog and the other working roles people confuse it with. A therapy dog is assessed to visit facilities by invitation; a service dog works in public by right, for one disabled handler, on tasks. The test exists to demonstrate that difference in behavior, which is why I think it is worth running even though nobody can make you. If that distinction is what brought you here, I have laid it out properly in service dog versus therapy dog.

And if your dog’s tasks are psychiatric rather than physical, run it anyway, and run it harder. A dog doing night terror work is doing something nobody in a grocery store can see, which means the only thing the public ever judges you on is the dog’s behavior in the aisle. Mr. Big’s tasks were largely invisible too. The composure was the whole public-facing job.

How to run one yourself

You do not need to hire anybody, and you do not need permission. Remember, the document does not say who may administer it.

A tripod and a smartphone is all the equipment this requires.

Take it in parts. Do the public sections in public and the private ones at home; there is no rule saying it happens in one afternoon. Then — and this is the part everybody skips —

If you are a veteran going through the VA for a dog, some of this will be handled for you and some of it will not, and it is worth knowing which is which before you assume. I have gone through that process in detail in how veterans get their own service dog. Even where a program is involved, the footage is still worth taking. It is your dog and your judgment that has to hold up in a doorway.

Go through the footage and honestly critique it. If you need more work, you need more work on something.

Being honest with yourself on video is harder than the test. You will want to explain away the one where he glanced at the kid. Do not.

And understand what going out in public does and does not do:

Going out in public is not going to change a deficit in training.

Exposure is not training. Exposure shows you what your training is worth. If the footage shows a gap, the answer is more work at home, not more hours in the mall.

What to do instead of, and as well as, a formal test

Train for daily life. The items on that test are your daily life — and daily life has considerably more in it than fourteen sections. Train for dead sure reliability and the test handles itself.

Having your service dog trained to the test threshold is the floor, not the ceiling.

The only published training benchmark, and it is voluntary

If you want a number to aim at, there is exactly one published and it is not from ADI. IAADP’s minimum training standards, which come down from ADI’s own mid-1990s originals, say:

• A minimum of 120 hours of training over six months or more.
• At least 30 of those hours out in public, in a range of settings.
• Not starting formal work before six months of age.
• Sit, stay, come, down and heel, plus a dropped-lead recall in a store.
• No biting, snapping, snarling, growling, lunging or barking while working.
• No soliciting food or petting.
• Ignores food on the floor.
• Identifiable trained tasks that relate to the handler’s disability.

Did that match what it took with Mr. Big? Not close. I did considerably more than thirty hours of public work before I felt totally comfortable with him out there. I was not counting, but months of the schedule I described above is not thirty hours.

Treat 120 and 30 as the floor of the floor. And note what IAADP itself says about failing their standards: doing so “does not mean the dog is not a legitimate assistance dog under the Americans with Disabilities Act.” Even the people publishing the benchmark will tell you the benchmark is not the law.

What a test cannot tell you

This is the honest limit and I would rather say it than sell you certainty.

A test will not tell you how your dog performs during a complete surprise or a catastrophic event. It cannot. Those conditions are not reproducible on a Tuesday with a tripod, and a dog who has never met a real emergency may not do well the first time one arrives.

What the test does tell you is that the dog reliably handles what comes up day to day, and that the two of you have built the foundation of trust and confidence to handle the rest when it lands.

That is not the same as a guarantee. It is a foundation. It is also the only honest thing anyone in this field can offer you, and anybody promising more than that is selling something.

The risk of deciding for yourself

Yes, there is one, and it is not the one people expect. The failure mode I actually see is not handlers overrating their dogs. It is handlers panicking.

Some of us think that if the dog is not perfect right out of the gate, something is wrong with the dog. Usually there is nothing wrong with the dog.

It is usually not a bad dog. It is a gap between handler and dog, or a gap in the training.

Gaps close. Mr. Big refused an escalator for weeks and then rode escalators for the rest of his working life. If I had read that refusal as a verdict on him instead of a job for me, I would have washed out a dog who went on to do everything I ever asked of him.

Why people think America needs a test

I get asked this constantly, and I do not think it is because of Australia.

People think a test should exist because it makes common sense. You do not want untrained dogs and untrained handlers out in public where real harm can land on somebody — on a stranger, on another dog, or on the team itself. That instinct is correct and I share it.

The problem is what happens next. I cannot see the US imposing a mandatory test and fee, because it would be unmanageable at the scale of a country this size, and because a fee on a disability right is its own kind of wrong. So the field is left to police itself, and an industry left to police itself usually does a mediocre job of it. Some handlers go to the ends of the earth to make sure they and their dog are safe. Others become exactly the problem everybody is worried about.

For completeness: a public access test genuinely is required by statute in Queensland, Australia, and British Columbia runs a roughly 40-task assessment with a fee in the region of $200. So if you are reading this from outside the United States, your answer may be different from mine, and none of the American law above applies to you. Start with does the ADA apply in other countries instead.

Where this leaves you

If you are training your own dog: run the test. Film it. Be hard on the footage. Fix what it shows you. Do it again in six months.

If you want a person to run it, Atlas at $125 or Medical Mutts at $300 are buying you real time from real evaluators, and both will tell you honestly what it does and does not mean.

If somebody offers you a certificate for $39 without watching your dog, you now know precisely what you are buying.

And if a store employee ever asks you for a test result, the answer is that no such requirement exists, and they are not permitted to ask. If it goes further than that, I have written up what to do when your service dog is denied access.

The test is worth taking because of what it tells you. The certificate is worthless because of what it tells everyone else, which is nothing.

Disclosure: the next link is an Amazon affiliate link. As an Amazon Associate I earn from qualifying purchases.

One thing that is worth buying, and it is not a certificate: a pack of ADA service dog information cards. They quote the two-questions rule in the Justice Department’s own words, and they are what I hand over when a conversation at a door has stopped going anywhere. They prove nothing about your dog, and they are not meant to.

Common questions

Is a public access test required by law?

No. No federal or state law in the United States requires a public access test, and no business may require you to produce one. Under the ADA, staff may ask only two questions: whether the dog is required because of a disability, and what work or task it has been trained to perform.

Who can administer a public access test?

Anyone. The ADI document does not name who may administer it. A trainer can run it, a friend can run it, or you can film it yourself with a phone and a tripod and assess the footage honestly.

Does passing a public access test certify my service dog?

No. ADI printed exactly that on the test itself — passing does not imply certification by ADI — and ADI states it does not certify individual assistance dogs at all. Nothing certifies a service dog in the United States, because no certification exists to be had.

How many training hours does a service dog need?

There is no legal minimum. The only published benchmark is IAADP’s voluntary standard of 120 hours over six months or more, of which at least 30 hours are outings in public. Most owner-trainers I know, myself included, do considerably more public work than that before they are comfortable.

Should I buy a service dog certificate or registration?

No. The Department of Justice states that such documents do not convey any rights under the ADA, and the companies selling them say the same thing in their own terms. A $39 certificate where nobody observes your dog documents your own opinion of your dog.

What happens if my service dog fails part of the test?

You train the part that failed and try again. A failure is information, not a verdict on the dog. Mine refused escalators for weeks and then used them for years.