Collecting Observational Data
Posted on January 15, 2008
Howard Abikoff (bio) talks about the difficulties of experiment design when working with children.
When we're thinking about developing a particular observational methodology, we consider a number of I think very important factors, and they all go into the development process of the measure. A simple one might be something like, how often are we going to have to collect information at baseline so that we are confident or reasonably confident that we have a stable estimate of functioning in that setting?
Well, in the case of children with ADHD, for example, we know from both clinical observation and data from multiple studies that the behavior of children who have ADHD is quite variable over time, both within a day and from day to day. Well knowing that immediately suggests that going in and collecting baseline observations just once will not be sufficient.
You may capture that child on a good day or a bad day. And what one has to think of then is, how often would one need to collect baseline data in order to composite it to yield what you hope is a reasonably stable estimate? So that's one issue we think about.
The other thing we think about, of course, is if we want to observe a child in order to collect information about certain aspects of functioning, well another issue is where do we want to do those observations? Now best case scenario is to go into the setting in the real world where the child spends time and to collect that information. So certainly in terms of classroom behavior it's very hard to collect that unless you go right into the classroom. Now that's very expensive; there's a lot of time and effort that goes into it.
But if your question is, I'm interested in how these youngsters interact with their parents and how their parents interact with them, well then the question is, where do we collect that information? Now we may want to go into the home, but doing so there creates problems because of the difficulty in controlling for setting parameters and contextual issues that may come up that the experimenter has little or no ability to control.
And if that happens, you're not quite sure whether or not the data you've collected are relevant beyond what just happened in the house at that time. So unless one is certain that that's something you can control for, and it's quite difficult to do that, it's not impossible, and there are folks who do do it, another way is to say, well perhaps I can get at parent-child interactions by bringing them in to the laboratory.
And there's been an incredible amount of work, and there have been some very creative attempts to develop analog situations in the laboratory where a parent and a child have an opportunity to interact in a setting in which there is a considerable amount of experimental control over key setting parameters.
Excerpted from interview with researcher in September 2007.
Please note that the feedback is viewed only by 4researchers staff and is not intended for communication with individual contributors.
Use the form below to submit feedback about this article. If you would like a response, please be sure to include your e-mail address.