Data driven instruction is the philosophy that data, about students, what they know and can do, who they are, and how they learn, ought to be driving the instructional practices of teachers and school leaders. Once upon a time, teachers’ observations, opinions of students’ abilities, and other anecdotal data was sufficient to drive instruction. While valuable, these subjective measures are full of potential error when it comes to who is, and isn’t, actually learning, and the underlying reasons why. Especially since educators should be using the information they have about students to CHANGE their instructional practices to meet the needs of students. Observations and anecdotal data simply may not capture enough information about students’ challenges to inform instructional changes. This is to say nothing of teachers’ biases (intentional or not) which may color their assessments of students’ abilities. So, we need quantifiable data about students (and teachers) in order to more accurately diagnose what, and where, instructional challenges exist.

So, what kind of data do we want ? Ideally, on any given assessment or assignment we would want to know the specific skills or standards on which students are struggling and the specific questions and questions types (multiple choice, matching, essay, etc.) on which students are struggling. This information would allow teachers to target specific skills and questions types. If students master a given skill when it is presented as a multiple choice question but struggle when presented as an essay, the teachers knows that she must provide students with more opportunities to practice essay-type questions.

For selected response questions, we would also want to know which incorrect answers students are selecting. This data would allow us to see if there are an patterns in students’ misconceptions, thereby allowing the teacher to pinpoint her remediation efforts.

Finally, we would want overall class averages, individual student averages, and we would also want to split students into performance quartiles (the bottom 25%, top 25%, and so on.) By grouping students into quartiles, the teacher is able to differentiate instruction for groups of students who perform similarly on a given assessment.

Given the above data, every teacher would be able to hone in individual students’ challenges, the challenges of groups of students, and the needs of the entire class.

So, why doesn’t this approach work ? Because, at present, processes for capturing and analyzing this data for every assessment either don’t exist or the turnaround time on the data is so long that the information is useless by the time the teacher (and students) receive it. To be fair, online testing platforms and electronic machines/software provide some of this data, but it’s tough to customize the reports to get to the kind of information teachers need. Some schools operate in areas where student performance is always great, so they don’t even bother with capturing this kind of data. Other schools have to be creative. There are schools that use paraprofessionals, volunteers, and clerks to capture this kind of data for teachers. These schools are incredibly fortunate, but they are not the norm.  Other schools expect teachers to use more of their own resources, most notably, their off contract time, to crunch these numbers. I was one of those teachers. And it sucked. I would bust my tail with students, give up my planning periods to have meetings about things that rarely led to student improvement, and spend countless hours after school or at home grading papers…and then spend more time creating my own systems in Excel for gathering the kind of data needed to make informed changes to my instructional practices.

So, what do we do ?

I think it’s important for school districts to either invest in, or develop, systems and processes for capturing the kind of data that teachers can use to make changes to their instructional practices. These systems should be tailored to elementary, middle, and high school use and should also be appropriate for capturing data for both selected and constructed response questions.

Once that happens, data driven instruction will work. Because teachers will have the data. Now. Districts have a significant responsibility in this equation, one that they’ve been shirking for far too long.

As always…thanks for reading and pardon the typos.


As far as easy to use data systems go, I really like the potential of the Quick Key App because of the turnaround time. Check it out and let me know what you think of it. I’ve used in the past and have loved it, but the last time I used it, it lacked a little bit of what I needed in terms of data analysis.

Why data driven instruction doesn’t work

Leave a Reply

%d bloggers like this: