 I think we'll start this out with the demo. We got the Hilbert chain as you're going to have eight hands. One hundred and five of us! You grabbed by the pumpkin full of her. We might be wrong. We're funny. I believe you've made it this far, but since you have, then we might as well talk about a completely different topic that we haven't talked about so far unless it was in a Penny Packers-Pettin video. So the trick is to stretch this one out to make it a little bit more informative than a Penny Packers-Pettin. But you're probably wondering what it is that we're talking about. It's an ABAP design. They're really popular. They've been in the field forever. In fact, you could argue that our entire field is based on ABAP designs. And for those of you that don't know what an ABAP design is, it's an A-B-A-B. So it's just easier. I get confused with letters, so just saying it as a word is much more fun. And now I can imagine going to ABAI and hearing people, look, I have an ABAP design. The first person that does that at ABAI, when I'm in the audience, is going to get a very big reinforcer. It won't be money because I'm broke. But it will be something else, probably just a clap. I said, well, I might even stand up and ovate you. Is that a word? Ovating? I guess it's a word. So ABAP designs, ABABs, right? Pretty straightforward, as you might imagine. It's really a repetition of two designs, an A-B and an A-B. Or it's an ABA, or it's an ABA design with an extra A-B, I mean a B. So we could get really confusing here. ABAB design. And I'm calling it that because I'm not going to get into the withdrawal versus reversal issue here yet. I will come back to that. So ABAB really is two designs. Well, one design repeated. A-B and an A-B. And you repeat it. Why? Replicability, of course. Because an A-B design by itself completely stinks and you can't draw any flipping conclusions at all. And in fact, what we call an A-B design, basically it's just known as a case study. So when we put them together, you start to get some cool things. Why is that? Well, because of the logic of the single-subject baseline. And as we just looked up on our channel, we've got about 14 minutes of discussion on single-subject logic and stability of baselines. So we don't need to worry about that here, but so you get your nice stable baseline and then you implement an intervention. So we do it over here, I guess. So you get your baseline and you get an intervention and hopefully behave it changes. Well, we're not completely sure if the intervention is what's causing the change now in the sorts of concomitant or confound or whatever variables you want to call them that's popping into play. So we reverse it. Oh gosh, we withdraw the intervention to see what happens to the behavior. And ideally, the behavior changes right at the same time that these intervention phases change, right? But in reality, it doesn't always change that quickly. There are some transition phases in here, which is why we wait till stability. And then we put that intervention back in place. So we do our A-B and we go, wow, we got results. So another A is like, hey, it's working. The behavior is going down or going up, whatever it is that was happening. And then you put your intervention back in place. As long as those things happen all at the same time, like as they seem to be related to the phase changes, then we draw the conclusions that these things work, right? That it is a strong, or it has high levels of independent, my gosh, internal validity, sorry. So it seems like these have high levels of internal validity if you do things correctly and if you see those particular effects at the time that you make the changes. Is it perfect? No, not at all. Like I said, those phase changes create transition points and those transitions can be difficult to interpret. So you need to make sure that you're attending to the data itself and allowing it to stabilize, right? Before you start switching conditions, I can't tell you how important that is because I can't tell you how important it is. It is the key of all of it, is the stability, right? I know you're probably wondering how about the predictability and the verification and replication and all that fun stuff. Well, baseline, you figure it out, it establishes something that you can then predict from and then you make your changes accordingly and then of course you get your replicability built into the ABAB design. So it's pretty cool, right? So a couple other things about this. Various texts go into different types of ABAB designs using different types of procedures to switch between, right? I'm not a big fan of getting into that nuance, why? Because once you understand the logic, then you don't have to pigeonhole yourself with a particular type of intervention to do for your withdrawal. You can do, sure you can do an NCR, sure you can do a DRO, sure you can do a DRA, DRI, sure you can do punishment, sure you can do extinction, you can do all of these different things. So don't limit yourself by just what's written down in the textbook understand that it's the logic now as I promised I was going to come back to this withdrawal versus literature withdrawal versus literature, whatever. So I'm having withdrawals from literature oh my gosh bring me the books now. So no, withdrawal versus reversal issue. So I think to be completely honest with you because I've been very saliently corrected on this in the past that a vast majority of the ABAB designs that we run into or run these days are really going to be called withdrawal designs. They're probably not reversal designs. Traditionally reversal designs have been referred to as designs where you're literally trying to reverse a behavior, choosing a different behavior to work with, right? So a DRO, DRI, a DRI type situation. So unless you're really reversing something then please just call it a withdrawal design. If you're removing a treatment from it you're just withdrawing that treatment. And thank you Ron and Nancy Martella for abundantly correcting me in front of all of my peers and colleagues one day. Because I joke about it but it was true. I went for a long time believing that they were exactly the same thing and like many people in the field we just call them reversal designs and it's taken a long time for me to fix that particular behavior. So please, yes, do focus them on withdrawal designs and don't worry too much about all the little details of the different ways that you can pull them off because it's really about stability and baseline, switch to an intervention, wait for stability, switch back to your baseline, wait for stability and then go back to your intervention, wait for stability. We're going to go over one really important one here really quickly which is permanent behavior. If you teach a permanent skill you're probably not going to be able to withdraw it. It's really like you can't. If I once I teach you to juggle I can't teach you to unjuggle. It's just not going to work. So things like permanent behavior changes we're not going to do those. We're not going to use ABAB designs in order to analyze or in order to study those. The other thing that pops into play is ethics. So if we end up with self injurious type situations we have to worry about those going back to baseline things and allowing baseline to stabilize and all that. You guys are already aware of that. I'm sure you've read all the textbooks and all the articles on ethics. So please be aware of that. But they are hands down pretty much the most powerful design type that you can come across. I've often argued that they are have the most internal validity of all the designs out there both for single subject and for group research. All of them are rehearsal. Enjoy your withdrawal designs and interpret them with vigor. Subscribe. Share.