Data collection is a big part of what we do. We are bound to make data-driven decisions and update or revise programs based on what the data tell us. However, it is possible to be so consumed with data that it interferes with learning. How well can a person engage and think on their feet when they’re preoccupied with data collection? Another downside of data is that they can hold a learner back from progressing. It may look like a student isn’t mastering any goals, when, in fact, a different type of data collection method may tell a different story.
We feel strongly that data should only support learning and never impede it. This is why it’s important to choose a method of data collection that allows us to monitor and view change, yet doesn’t hold the learner back…. AND one that permits the instructor to be present in the moment, engage with their learner, and do good teaching! Students will learn more from good teaching than from good data.
We have a cheat sheet of different data collection methods available for you to download below!
The Best Data Collection Methods
There are many ways to collect data. Here are a few of our favorites:
Rating scale data
A rating scale describes behavior along a continuum. Although rating scale data can be subjective, they are a great tool to use during play with early learners. They can describe the level of independence (or resistance) for an activity without impeding teaching. During the time when the instructor and the student are getting to know each other, relationship building and engagement (a.k.a. pairing) is first and foremost. A clipboard or electronic data collection may ruin the quality of that interaction. Rating scale data can outline some goals for the instructor to work on and then the child can be rated at the end of the session, based on a continuum of how well they did that day. The scores are then graphed. Check out more about rating scale data sheets and download a free copy.
A rating scale is also a great measurement tool for group-based learning and NET (natural environment teaching). The therapist can follow goals and teach effectively without being tied to a clipboard, clicker, or pencil. The quick rating scale allows us to track progress without it being too cumbersome.
Frequency and Duration
Frequency refers to the rate of response. In other words, how many times are you seeing the occurrence of a particular target? Tally counters are a great way to collect frequency data. They are not cumbersome! Requesting and spontaneous language are some programs where we may choose a frequency count over any other type of data collection.
Duration refers to how long the behavior persists. We often take duration data when we want to know how long it takes students to complete specific tasks. We also take duration data when analyzing negative behavior, like tantrums. How long do they persist?
An efficient way to track negative behavior is to use a partial interval recording. By checking off the minutes/times that negative behavior is occurring, we get information about both duration and frequency. This can then be graphed so that we can see trends in behavior.
Probe data means that we take yes/no data on the first trial of a program and then throw our pencils down and teach! Probes can also be done less often if in a classroom situation (e.g., think spelling test), or in the community. Teaching should still happen after the probe (especially if it was incorrect!). Mastery may be something like “3 correct responses across 2 different people”. Probe data can be great for progressing students who learn quickly through programs. They may only need a few exposures of the material. For example, a probe data sheet can be used with a program like, “Expressive Labels” so that multiple targets can be run simultaneously and the student can advance through them quickly. Caution: the downside of probe data collection is that instructors might not run as many teaching trials as they should if there is no data to record after the first response.
Taking trial-by-trial data means that we record data for each trial run (usually at least 10) and then get a percentage of correct responses. Mastery is usually something like “80% or more for 2 sessions, across 2 people”. Some children need more exposure, and repetition is good for their learning. Collecting trial-by-trial data encourages instructors to run at least 10 trials. Graphing the percentage tells us more information than probe data. It is also more objective than rating scale data. By graphing the percentage, we can see if the curve is increasing or decreasing or if there are any other notable changes.
As Behavior Analysts, our skills lie in taking the strategies we’ve mastered and applying them differently to each learner. Make sure that the data collection method you choose is helping, not hindering, your student’s progress.
Make sure you download our data collection methods cheat sheet!