On the Reformation of ABA

Back in December (2025), I encountered a content creator who appears to be a twenty-ish autistic advocate. Much like myself at their age, they're passionate about autism acceptance and various related topics. Unfortunately, one of those topics is ABA reform. To create a new, modern ABA that is helpful to autistic people. 

Over the years, I've encountered a few autistic advocates for ABA, who think that ABA helped them and/or that the ABA industry can be changed for the better and not be abusive. I don't know if I will change any minds, but I want them to read and think.

Let me put it plainly: I don't think it's possible to reform ABA or its industry. Even if there were any potential positives to ABA, I think that it's impossible to separate it from the harmful aspects. Other people have covered these topics more extensively that I possibly could, so I'm including some links that I hope are useful at the end of this post. 

But let me approach this differently: ABA is not backed by science. 

Yes, there are a lot of studies that argue that ABA works, but they do not meet the criteria to be considered gold standard science.

To summarize, to be considered good science, a study needs to have a clear research question, be designed appropriately with large enough sample size and control groups, and be reproduced by other scientists. It also needs to be transparent about any conflict of interests, sources of funding, methods used, the data, and be honest about the limitations of the study. It then needs to checked by peer review, in other words, be reviewed by other scientists not related to the study. 

Double-blind studies are often the best methods to reduce biased results in studies, but I do admit it is very hard to do in social sciences. It also takes a lot of time and resources (including money) to do good studies. 

So what I focus on is transparency, sample size, and limitations of study.  

There has been increased criticism to undisclosed conflicts of interest in pro-ABA research studies, as well as criticism over the ethics. I'm not going to go into these points, but it is not a good indication of reliability if researchers do not fully disclose their conflicts of interest. It gives doubt to their biases and to their results. If they are not being honest about their conflicts of interest, it is possible that they are not being honest about other things in their research.

My particular focus in on the sample size and limitations. From what I understand, a larger sample study gives better statistically significant results; in other words, it makes it less likely that the results are flukes. Currently, there are very little large scale studies that support ABA. In fact, the large study conducted by the U.S. Department of Defense reported that most participants saw no improvements over a year of ABA. 

Nevertheless, in order to have significant results, a study needs a large sample size and at least one control group. In other words, there needs to be a group that the study compares against. So if there is a group that receiving ABA instruction to complete a task, there needs to be a group that is receiving non-ABA instruction to complete the same task. 

In my idealized world, a study would, at least, have both autistic and non-autistic groups, divided into those receiving ABA instruction, and those not. So four study groups of participants and instructors. Instructors would be blind as to what group they are instructing and how they are instructing, and participants would be blind as to whether they are receiving ABA or not-ABA instruction. So it would be a double-blind study. This way the researchers running the study would not have their biases distort the data. (There are probably other factors concerning a large scale study that I'm not mentioning, this is just a simplified scenario.) 

I will admit, it is difficult to secure the funding to run such a large study, however necessary in order to have significantly reliable data. Still, from what I have seen of pro-ABA studies, very few of them have a large sample size, do not have a control group, and present very weak data to their efficiency. Most studies set out to prove that they can teach a skill or task using ABA.  

Which brings me to limitations. There's always going to be limitations in studies; there's only so much one can do with whatever funding secured. What often gets omitted is whether the task or skill being taught to participants is something they want or need to learn, whether they retain what they learn in the long term (in years, not months), and whether there are negative effects of the instruction they received. 

What good is a treatment or therapy if it causes more harm than good? The pro-ABA studies do not look at any future results, both good and bad. They do not study whether receiving the ABA instruction was beneficial or not for the participants. The only studies that do indicate that ABA causes PTSD in autistic people. 

Which, of course, I expect pro-ABA advocates to circle back and say that is why they want to reform the industry. But what good is reforming the industry when the practice itself is the harm? Especially when there is very weak data indicating that it does any good in the first place. 

I don't expect to completely change pro-ABA advocates' minds. But, if you truly believe that ABA is beneficial and the industry can be reformed, I ask for you to advocate for better studies. The current data is very weak and inconclusive, while the bad behaviour of researchers undermine their results. And while I don't believe ABA will ever be beneficial to anyone, I do think that autistic people deserve to have good science-based evidence to back their advocacy. So if you think that ABA is valuable, then having good scientific studies can only strengthen your advocacy. 

Personally, I think that it is more beneficial to focus on autistic led goals, such expanding healthcare coverage to things other than ABA and providing individual-directed support across the lifespan. To putting research into supports and services that do not cause harm, and instead seeks to repair harm done. 

Other References

The Troubled History of ABA Therapy

Amythest Schaber's ABA masterpost

America’s Most Popular Autism Therapy May Not Work — and May Seriously Harm Patients’ Mental Health

5 Important Reasons Even New ABA is Problematic

A4A Why We Oppose ABA in All Forms

Behavioural Modification Therapy Does Not Work

The Great Big ABA Opposition Resource List

Why No Autistic Child Should Be in ABA Therapy