Image via Wikipedia
I spent all day yesterday working on a continuing education class to maintain my behavior analysis certification. The class is about the effective uses of punishment, and I’m sort of hoping for some feedback on one of the essays I had to write. Whether you’re down with the behavior analysis or not, hopefully you’ll find it interesting! …Enjoy… (well, as much as you can enjoy an essay about time-outs, I guess)…
(This essay a response to a prompt asking how one should analyze the information given in research studies, specifically how to know which interventions will work and which ones won’t).
If you’re writing a paper about the correlation between, say, hitting and time-outs, chances are there’s a research study out there somewhere proving a substantial, clinically significant link between the two. More than likely there is also a research study proving exactly the opposite. As consumers in general, and specifically as behavior analysts, it is our job to figure out which research studies successfully prove what they say they do, and which do not. Failing to consider behavioral interventions on a case-by-case basis not only promotes misleading results, but can also negatively influence future intervention techniques.
It is important when analyzing a research study to look at what the author DOES do as well as what he DOESN’T do. For example, let’s discuss an imaginary article stating that time-out works. If an analyst takes excellent intervention phase data but relies on anecdotal evidence for baseline, we must call into question how accurate his baseline data is. If he, say, correctly defines the target behavior, but fails to do a functional analysis of it, how can we know if his intervention actually influenced the target behavior, or if he just got lucky? If all the information you have is that “when Johnny hits he gets a time-out”, theoretically you’d have to put him in time-out for hitting a baseball, and, since time-out isn’t clearly defined, you could give a five second time-out or a five-hour time out, and both would be recorded the same way. Life, as it were, is in the details. So lets say that according to this hypothetical research study, time-outs do, in fact, work. But after looking a bit closer, we should actually say something like, “Time-outs decrease the frequency of Johnny’s hitting behavior when the behavior is maintained by attention and when the time-outs are implemented consistently for a period of ten minutes each.” Did the author of this study prove his correlation? Yes. In this particular case, time-out works. Does this correlation automatically lend itself to a general rule? Absolutely not.
The same can be said of analysts who argue against certain correlations. Those who say time-out doesn’t work, for example, probably found that in their particular situation it didn’t. However, a wise behavior analyst must understand the plethora of factors that can, and do change in each situation. The value of the rule in general is never as potent as the value of the specific behavior. Time-out, then, might not work for Sally’s hitting behavior because her behavior has a different function. Sally doesn’t hit for the same reason Johnny does, therefore the same intervention probably isn’t going to work to decrease Sally’s behavior. Maybe time-outs didn’t work for Sally because Sally’s dad is a big softie and he let her get up every time she started to cry. Maybe time-outs didn’t work for Sally because she enjoys quiet time. The analyst who wrote Sally’s case was absolutely correct in stating that time-out did not work for Sally. Can we, as behavior analysts extrapolate that research study to mean that time-outs never work? Again, the answer is no.
So, time-out works, but not always. Published research studies generally mean what they say they mean, just not exactly. And you can trust the research that’s published, except when you can’t. Essentially, it is our job as behavior analysts to consider behavioral interventions, those in published studies and those we encounter in the field, on a case-by-case basis. We must understand that correlation is not the same as cause and effect. We must use our critical thinking skills to recognize that time-out isn’t always a punisher, just as candy isn’t always a reinforcer. We must analyze not only behaviors, but also the research completed by our peers, and use our best judgment to decide which interventions will be the most effective for behavior change in each specific client. If we fail to use all the information we have, we not only do a disservice to our profession, but we also fail to provide the best intervention to our clients.
Alright, well, after reading this again I have a few questions/thoughts:
1. Wow, I really need to brush up on my comma usage skills.
2. I’m really, really hating my conclusion. It’s just… forced, or something.
3. I intentionally took a casual approach to the language, mostly because I hate writing sentences where the subject is the word “one.” However, I’m now thinking it might be a bit too casual.
4. I also intentionally did not use gender neutral language. In my mind, the author was a man. That’s why I used the pronoun “he.” “(s)he” is not a word, and it hurts my soul when people actually use it. BAH.
5. How are my tenses, mom?
6. This is a draft, so there are still things I know I need to work on (getting all that nasty passive voice out, for starters…)
I’d love your thoughts, friends! })i({
**UPDATE** I finished the essay this morning; if anyone’s interested in the final draft just shoot me an e-mail! (8.18.10)
What Not To Crochet
A Happy One 
I thought your comma usage was better than most. I personally believe society should abandon the idea of gender neutral writing. I was drafting a court opinion recently and having to say “he or she” repeatedly breaks up the flow of a sentence and also is distracting. I found a way to use “he” as well because the party was a man, but I had to rework the entire paragraph.