ABA Industry Research is Highly Problematic

Not only are studies extremely weak with pervasive COIs, now they’re being written by AI…

Created by author

In March 2023, an ABA practitioner named Sara Gershfeld Litvak published an article in a (supposed) peer-reviewed journal called the International Electronic Journal of Elementary Education claiming ABA-based therapy is “the only non-psychotropic scientifically validated treatment for children with autism” and that there is a “national shortage of ABA providers in autism treatment”.

The citation used to support the first claim was written by one author, Dr. Brian Reichow, not multiple authors as Gershfeld Litvak’s citation implied (as it used et al.) In his 2012 article, Reichow expressed concerns about publication bias because journals related to the ABA industry tend to only publish “successful” studies, meaning studies which offer findings that the industry already wants to convey to the public.

Reichow also found discrepancies in reporting of both child and treatment characteristics. Those who have endured statistics courses might remember this means the data-gathering lacked inter-rater reliability. For example, different data-reporters gave different ratings of the same behaviours. (The level of inter-rater reliability “represents the extent to which the data collected in the study are correct representations of the variables measured”).

The citation given for the claim that there is a national shortage of ABA providers in autism treatment was missing from the references listed at the end of the article.

Screen shots by author

When I tried to find it on my own…

It doesn’t exist.

Not only is it not listed in the references section, this phantom reference is completely made up. Searches for Pennington & Fedele (2020) or any variations on the argument made in the article turn up nothing. In fact, about seven months later, the article was retracted by the journal’s editors.

“An article written by Gershfeld-Litvak was retracted at our request due to the use of Artificial Intelligence (AI) that led to numerous inaccuracies within the reference and the body of the paper.”