In science fiction, an engineer will frequently tell an authority figure how long it will take to do repairs after a disaster: “It will take at least an hour.” Then the authority figure says, “You have ten minutes,” and somehow the engineer finishes the task in time.
The thing is, when you have a complex problem, no one knows exactly how long it will take. Since there is substantial variability between different children with autism it can make this problem more challenging. In addition, most of us are generally bad at estimating time. In most cases, we don’t have a science-based way to make a good prediction of how long it is going to take to solve a complex problem. Even when we have data like that, they are limited and typically don’t include the full solution to the problem like ensuring generalization of the skills and maintenance. Also, it is rarely under practical conditions like a school with many different goals being implemented over a relatively long time.
Even though we don’t have great science-based ways to predict how long things are going to take, there is an expectation that we can do that. In addition, people assume that those estimates are padded; if the captain demands it, we can increase how fast we solve those problems. This leads to several concerning issues like making it difficult to judge the effectiveness of a program.
If a client starts with a baseline of 10 and increases to 25 in one year, we should evaluate that. Unfortunately, we don’t measure that way. Instead, we try to predict how much progress the client “should” make in one year. So, if we predicted that the client’s data would increase from 10 to 20, people will likely judge the client as having made excellent progress. On the other hand, if the goal was 50, we say the client didn’t do very well. The level of progress didn’t change. Just the comparison to our initial prediction.
The internet is uncertain of who originally said it, but “predictions are generally very difficult to make–especially about the future.” In the field of behavior analysis, we can often predict accurately whether a client is likely to be successful or not successful with a particular behavior change program. We aren’t very good at predicting how much will be able to be accomplished in a particular amount of time. This is especially true when there are lots of competing goals happening at the same time.
We are much better at making predictions when the behavior changes are small and the expected time frame is short. Ideally, we work as much as possible in that realm. Set small goals that can be met quickly, and frequently write new ones. That method of working might not be possible in every environment, but if you work at it creatively, you can do this in more places than you might think.