Author: Barry Morgenstern

Many programs working with children with autism have huge problems with turnover. Now there are many reasons for turnover- and they will differ as schools and other programs will have their own individualized issues. The usual suspects- pay, benefits, commute, schedule, advancement opportunities, and quality of managers- are, of course, very important. If the school or program doesn’t have at least minimally acceptable levels of those six components the program will likely have ongoing problems with turnover.

A typical thought process among behavior analysts often goes like this: Well, there is nothing we can do about reimbursement rates, the health insurance, the fact that the therapist lives an hour away, the school schedule, or the fact that the family needs someone on weekends. There are so many things that are simply out of the typical behavior analyst’s control that there is sometimes a hopelessness about improving retention.

As a behavior analyst, there is one thing that is largely under your control. And if you do it well it will drastically reduce turnover. In my experience, a simple thing you can do that substantially improves retention is having high quality programs.

High quality programs lead to many staff reinforcers. First, it is exciting to come to work when you see kids making great progress. Second, if the program is high quality the staff feel like they are learning which is often highly motivating. Third, the staff feel a sense of satisfaction when through their own efforts they are seeing huge improvements in children’s lives. Fourth, time during the day starts to feel like it is moving quickly. They are enjoying themselves so much they forget they are working. When this is happening, turnover is lower. Some staff will get a 2^nd job if the pay at your agency isn’t high enough. They will deal with the long commute- if the child they are working with is making exciting progress. They might even deal with a manager they don’t like- if the person is really good and helping them achieve excellent progress with kids. (Sorry to all the folks that had to deal with Barry from the Bronx in the early years.) Will everyone stay? Of course not. But we underestimate the reinforcers that come to staff with high quality programming.

Low program quality leads to many staff punishers. The job starts to feel boring. Staff begin to not really try to teach and just go through the motions of running programs. Looking at the time to see how much time you have left in the day becomes common. Under these circumstances people will jump at any opportunity to get out. We dramatically underestimate how much low program quality impacts turnover. Will everyone leave? Of course, not. Some people will stick around for your other six components. But if they are sticking around just for pay or health benefits, you probably don’t want those people anyway.

I have briefly and not-very-seriously studied a variety of martial arts. But as I started to study behavior analysis, I had learned enough about the martial arts to become very skeptical that most martial arts students would be successful in an actual situation that required self-defense. That’s because in martial arts class, you study under very contrived situations (e.g., put your arm like this, OK now grab my wrist). Now, after a long period of training, it can look really impressive. There might be many benefits (e.g., social, health, discipline, focus). But would the skills a student develops in a typical martial arts class work in a real-life self-defense situation? In my view, it’s unlikely but pretty hard to do the research or get that data.

Now in behavior analysis, pretty much the only thing we are interested in is getting the behavior to work well for the client in real world situations. In some programs we use practices that are very contrived (like a martial arts class):

• Get the child’s attention first
• Ask what do you want to work for?
• Make sure you speak clearly
• Provide tokens contingent on correct responses
• Use an errorless teaching procedure
• Contrive the opportunity to mand
• And on, and on, and on

Now there is nothing wrong with any of the above practices as a training technique. But I worry that we are sometimes satisfied with the impressive looking demonstration (like a martial arts class) without doing much fading of the types of procedures used above. Or actually checking if the skills we taught work in the real world.

We are different than a martial arts class. We can go and see if the child is requesting with mom, having conversations with peers, or learning cognitive and academic skills in the general education classroom easily (well relatively easily compared to seeing whether self-defense skills are effective in the real world!)

It is heartbreaking how often I’ve seen skills that looked amazing under specialized sets of conditions but don’t work under real world conditions. It is easy to fool yourself that everything is wonderful if you never leave the contrived situation. Even worse, when it doesn’t work in the real world after we have spent enormous time, effort, and emotional energy, we get upset. There is a tendency to blame others (e.g., the parents wouldn’t implement it, the school won’t support the program).

In order to Poogi, you have to have the right criterion for success. That criterion is always the question: Does this skill work under real world conditions? If it doesn’t work (or worse you haven’t checked), you aren’t finished yet.

Many studies suggest that training peers to interact with children with autism can be an effective strategy to teach social skills. Training peers has a lot of potential for many reasons: It can create social motivation as well as many opportunities to practice, it sets up a realistic context to practice, and it can produce natural reinforcers. Sometimes these skills generalize and the children with autism start using their new social skills with other children too.

But sometimes the skills don’t generalize and other interventions are needed. The key risk is that sometimes we can be easily fooled by data that looks like the child is making amazing progress. The data may look great as long as the child is interacting with trained peers. The key point is you don’t really know if you have anything until you bring in different, untrained peers. In other words, sometimes you can’t be sure whose behavior has changed. Was it the typically developing child that was trained to prompt and interact, or did the child with autism develop real skills? Without natural contingencies the behavior is unlikely to maintain.

The same logic can easily be applied to typically developing children who were not specifically trained, but just naturally prompt and reinforce appropriate behaviors. These children can help produce amazing benefits, just be careful not to be fooled those data. It may be that real skills have developed, but you don’t know until you check. As soon as you hear someone say, “S/He is amazing with him” your radar should go up. Interactions like these are often a great start to helping developing social skills- just don’t get fooled that they produced more improvement than they really did.

Being brutally honest with how much progress you have actually made is crucial to the Poogi.

I went to the dentist and he said, “Barry everything looks good, but er um, er – Can I ask you a question?”

“Sure,” I said.

“Have you been punched in the face?”

Of course. Lots of times. Occupational hazard.

My dentist went on to tell me that he couldn’t tell how long ago it might have happened, but definitely looks like a punch. Probably have to replace that tooth. Expensive. Yuck.

If you have been doing this awhile, you have probably had feces thrown at you, been punched, bitten, scratched, kicked, spit on, or had other pleasant work-day experiences. Now, as our treatments have been Poogi-ing over the last 25 years, this happens less and less. But episodes like this are still inevitable from time to time. In all likelihood you handle those punches in the face well. You can remain calm, act professionally, ensure safety, and provide effective treatment. But if my experience is common, there is another kind of punch in the face that you probably don’t handle nearly as well: Criticism.

Despite being models of professionalism in dramatic situations, let someone criticize the prompting method or the data collection procedure and watch behavior analysts freak out and engage in absurd unprofessional behaviors.

An important behavior if you really want to be on a POOGI is to carefully listen to criticism and use it to improve. Sure, sometimes the critic is wrong and you want to ignore that. But criticism is often a gift. If you really care about the POOGI, you should reinforce when people criticize. That’s how it works. You have to be able to take both types of punches.

One thing that I learned from the book The Goal is it’s possible, even probable that when you have contingencies attached to data, people will engage in behaviors that are not necessarily in the best interest of the client.

For example, one teacher told me that he worked in a program where the number of trials per day was publicly posted. Now, in general, he thought this intervention had positive effects. For the most part, more well-delivered trials will promote more learning. But this intervention did produce a nasty side effect. Specifically, there was great resistance from the staff when programs were suggested that made it difficult to get a high rate of trials.

So in this program everyone was happy to work on reading that required frequent responses, math problems, and discrete trial instruction. But what about changes a child might need to make – that are absolutely critical- but don’t produce a high rate of trials? Mands for Information? Problem Solving? No one wants to work on those programs; their trial-rate graph will look terrible.

A cardinal rule that I try to follow is: If we implement this measurement, will it encourage everyone to engage in behaviors that are in the best interest of the client? Or ask, “Is there any way staff can show improvement in this measurement while doing something that is not in the client’s best interest?” If so, either don’t implement the measurement or add other procedures / measurements to attempt to prevent the negative effects.