Student assessment data has many utilities – it helps to describe the effectiveness of a school in terms of moving students forward, it offers great information to help guide the work of teachers and administrators, and it serves as an important early warning system for at-risk kids. How could something as simple as a chart with some scores tell us about the deeper needs our kids may have? That’s a fair question, and the answer – some behind the scenes data work!
One specific type of work starts with a process called segmenting, where we slice our information pie in different ways to understand kids better. Imagine you are a 6th grade Language Arts teacher, for instance. It’s the beginning of the school year, and you want to see what knowledge your students kept over summer and what they lost to the break. You have an assessment ready to go. You can administer it and get a feel for the classroom overall, but you can also summarize the results in a different way to get an even clearer picture of individual student situations.
Say in this instance that you have a pretty diverse class: there are a couple of kids who were home schooled and may have some reading risks, there are a few kids who don’t speak English at home, and there are a handful of kids going through the evaluation process for special education services.
With all of this in mind, it probably makes sense to go from this sort of a roster:
To this sort of roster:
By segmenting the list, the new roster allows you to determine how well kids in different groups are doing. You get a baseline for each segment, and you gain the ability to use more specific interventions instead of a blanket approach. As educators, segmentation can potentially offer powerful insights into what will happen in the future.
Segmentation can also be used on a larger scale, such as for all of the students in the school or district. We use statistical analysis to ask questions about the relationships between different segments and whatever we are interested in – in this case – a score. And what if I told you we could use the score to figure out what’s going on in another column of data we care about? As luck would have it, we can!
In a school I led, we looked at student data at several levels. We found that we were able to attract at-risk kids, but we weren’t always able to keep them. When we surveyed the kids who’d fallen off, we found some common problems. We developed a way to ask new students about these issues up front to help us intervene before it was too late. Our strategy involved an interview by a team member and a form that collected data about life variables and then blended that with assessment data to create a “risk score.” Once we got our risk evaluation program off the ground, we were able to predict student risk factors and cut our student churn down to 20% of what it was before we started.
For example, Ronnie (fictitious name), 18 years old, came to us and didn’t seem to have any risk factors. However, we noticed that he changed his address nearly every month. We got back a score on an assessment that didn’t seem to fit his capability or his past scores. A team member looked in his file for the interview form, didn’t find it, and re-administered it. We learned that Ronnie was homeless and were able to connect him to our school social worker to get help. As his life became more stable, his performance improved, as well. Had we not known that the achievement data had something to do with the risk score, we wouldn’t have known how to help him.
This served our school as an effective way to spot early warning signs for at-risk kids. In an upcoming post, I’ll talk about strategies that work with assessment data to identify and change unwanted behaviors. In the meantime, if you segment data to help kids at risk, tell us about it on Facebook or Twitter (@Assess2Learn).
This is the fourth blog in Dr. Jordan Argus’ series, School Rx. You can find his previous posts here.
Dr. Argus is a regular contributor to our blog and is an education advocate, speaker and consultant. Currently, he is an adjunct professor at The Chicago School of Professional Psychology.