The Matt Harrington Show

Journal review, extinction shaping, and why big box ABA is struggling - I read a study on masked visual analysis (MVA) versus traditional visual analysis in multiple baseline designs. Two methods only agreed with each other on 75% of graphs, so I dug into why. - My take: MVA exists because statistical analysis used to be too hard for clinicians to run. That barrier is mostly gone now. In five years I think insurances will want numbers, not graphs. - A student asked how to operationally define a kid shutting down. I said look at what they are not responding to and whether it causes harm. If it does not cause harm, define the actual behavior you want instead, like getting up from a task. - On tolerance building, I walked through shaping with extinction on small chunks of time. The key question is what you do after a kid taps out. You back off a couple seconds, get a few wins, then push forward again instead of jumping straight back to the old target. - On automatically maintained behavior, I explained how I treated spitting with competing stimuli and an enriched environment, then faded access with a multiple schedule. I gave a test for telling competing stimuli from a true functional match: does the original behavior come right back once the item is removed. - Someone asked if we should require student analysts to take a counseling course. I do not think the BACB adds much more coursework soon since it would slow down how many BCBAs get certified, and there is already huge demand for BCBAs. - A BCBA vented about non-BCBAs telling BCBAs how to do their job at big companies. I broke down why ABA businesses are splitting into three types right now: BCBA owned shops losing ground to competition, private equity groups with money but shaky clinical quality, and investor backed groups trying to hold both together. - My honest opinion is the businesses winning right now are run by a clinician or clinician adjacent operator who reinvests in clinical quality first and treats happy clients as the real marketing engine.

Journal review, RBT motivation, and a hard aggression case - I broke down a translational study on procedural fidelity errors that happen after a skill is already mastered, not during learning. 24 undergrads mastered a task on a computer, then researchers started feeding them wrong feedback on purpose. - The results were messy. About half the participants kept answering correctly even after errors were introduced. The other half dropped in accuracy. Two totally different patterns in the same study. - That split matters. The whole study assumed correct or incorrect feedback was acting as the reinforcer. Looking at the graphs, that assumption only held for about half the people. Some just kept giving the right answer no matter what the screen told them. - I said what I think plainly. This read like a thesis requirement more than a study built to move the field forward. Nothing wrong with that, I did the same thing for my own master's degree. - On the RBT question, I don't think pumping people up before session is the move. I'd rather know what each RBT actually cares about and explain the plan in language tied to that. Some want to hear about client outcomes, some want to hear about the clinical reasoning. Use the one that lands for that person. - On staff saying "no" or "we don't hit," I pushed back on the idea that this creates a new function for the behavior. That's close to what happens in a functional analysis and it rarely creates a function. The real question is whether that response fits the plan you're actually running. - I walked through a tough case, a 12 year old with aggression maintained by access, where the adults in the room can't withhold items. Straight extinction won't work there because the reinforcer can't actually be removed. The fix is finding the response class hierarchy, reinforcing a smaller precursor behavior instead of the full blown aggression, then teaching toleration in small steps and fading the wait time up. - FCT alone doesn't solve a case like that. Teaching the ask is the easy part. Teaching the kid to tolerate not getting it right away is the hard part, and that's where the real work is.