Data Driven Instruction

The new buzz word in education is “data driven instruction”, but, does this really work? The concept behind data driven instruction is simple. You teach a topic over the course of several days. You give the students a test based on the topic you just taught. Each test question is tied to a specific benchmark or standard from the topic you covered. When you get the results of the test (Of course which you could only get using the complicated software/scanner program that the county paid millions of dollars for) you can see exactly which benchmarks each individual student did well on, and which they did poorly on. You can even get averages for the whole class. So, if for example 80% of the class did well on a certain benchmark, you can move on and say that the students for the most part got it. However, if the reverse is true and only 20% of the students did well on a specific benchmark, you know that that particular benchmark has to be retaught or reviewed.

One of the problems with data driven instruction is that there has never been any kind of statistic showing that this method of teaching improves scores or makes any significant difference in student achievement. I have an acquaintance who teaches AP physics at the high school level, and she uses a teaching technique that was developed specifically for that class known as modeling. The idea behind it is that instead of the teacher lecturing physics concepts to the students, the students do hands on experiments and develop results themselves that help them create an understanding of the concepts behind the experiments they just did. This teacher periodically collects data about student learning, and sends it to the university which then compares it to data from classrooms with a more traditional learning environment. This team at the university is trying to ascertain whether or not the modeling technique actually makes a difference. After their several year study is over, they will have concrete numbers to know whether or not it actually is a superior teaching method. By contrast, we are being forced down this path of data driven instruction across the board without anybody having done any studies on it.

A second problem with data driven instruction is that there are several different types of questions that can be posed for every benchmark being tested. Assume that a student gets a question wrong on the test. Is this really because they don’t understand that benchmark? Or, could it be that they just don’t know the answer to that particular question, but if you were to ask them five other questions pertaining to that benchmark they would get all of them right? The data also doesn’t take into account student guessing. In order to get the data using this particular computer program that the district demands we use, all the tests have to be multiple choice. What if on a fifty question test, the students guessed on ten of the questions and got lucky, getting eight out of those ten correct? The data will tell us that the student is actually doing better than what they really are.

Also we have to take into account that there is a certain amount of material that has to be covered in a year. Assume you teach a class of low performing students who traditionally fail their tests. Are you going to halt your progress and re-teach a whole section because most of your students failed it? This would only make sense if you were teaching a subject, like a foreign language or like math, in which each new lesson builds upon the knowledge of the old. Otherwise, a teacher simply does not have the time to re-teach every lesson the students do poorly on. The most we can do is go over the test so that the students are aware of the questions they missed and become familiar with our testing format, and then move on. Otherwise by the end of the year we would not have gotten very far at all, and it’s a lot worse to end a course where students have gained mastery over 25% of the material but have not been exposed to the other 75% than it is to have covered 100% of the material, even if students failed a few tests here and there. At least they were exposed to the information, and next time they see it (because for most courses they certainly will see it again) they will have a rudimentary understanding of the information or time period.

Another interesting point; the other day I looked up the standardized test course of a particular class of students out of curiosity, because the administration seems to believe it is of such importance. I looked it up and saw that in that particular class I had most of my students at level 1 or 2, but I did have a few level 3’s and two level 4’s. What I found was that the two students that were of the highest level were actually the two worst students in the class. One student has a 0.0 GPA as of my checking, while the other student is of the type that can’t shut up long enough to learn anything and is constantly distracting and interrupting the lesson. Meanwhile, the brightest student I have in my class, a little boy who is actually been promoted a grade above where his age says he should be, was one of my level ones. This level 1 student answers my questions in class in such a way that shows he has a deep understanding of the material and thinks about it critically. The bottom line is that students should not be judged by their standardized test scores, because they do not define the student. Like a colleague of mine recently told me, “I was not a good test taker when I was in school; I couldn’t pass the SAT’s if my life depended on it! But today I have two Phd’s.”