by Ted Murcray
The semester is over and the grades are done. Grading is my least favorite of the teaching tasks because it seems so finite – so immovable. I prefer the teaching moments when I am working directly with students.
I take a look back over the semester and think about a particularly well-done lecture where the students responded well. Maybe I spend time reflecting on a well-designed discussion session that engaged students in an important dialogue. Personally, I hold onto these moments because they feel like clear markers that my teaching matters. What motivates me – far more than grading – is the chance to feel like my students are learning the material and skills that I care so much about.
That’s why I was so excited to read Fischer, Frye, and Hattie (2016) describe how they use effect sizes to determine whether students are learning. Instead of focusing on proficiency numbers (how many got an A, how many got a B), Fischer et. al. determine whether or not students learned by comparing pre-post measures and calculating an effect size for the class as well as for students individually.
Fischer et al. (2016) believe that teachers can determine whether or not students learned in their classrooms by comparing the effect size of a pre-post exam against the hinge-point of .40. Hattie (2009) set the hinge-point of .40 when he was working on an extensive meta-analysis of educational research – a project that took 15 years to complete. He claimed that .40 indicates that a strategy or tool (the educational piece being measured) is visibly or demonstrably different from the norm (read his book, Visible Learning, 2009 for more on that).
I started to employ this method on a few key indicators in one of my classes. One was a comfort survey, which I will talk more about in another post. Another was on a set of key information for my class. I created a quick quiz, which I administered at the beginning of a class session and then at the end of the same class session. The starting average score was a 1.35 out of a possible 10. At the end of the class session, the same students earned an average score of 5.8 on the same quiz. The data had an effect size of 1.86, which is well above the hinge-point of .40.
Although a final score of 58% percent is not what I hope to see on a quiz, knowing that the average score jumped up 40 percentage points is promising. I will talk more about what happened next in another post.
There are definitely limitations to using Effect Size as a measure of learning, and there are some reasons to think carefully as I analyze the results of this particular quiz, but I am encouraged to note that there is one way to get some data around growth in learning within my class. I want to know that students are leaving my class knowing more than they did when they walked into my class, and effect sizes are one way that I can get that data.
Fischer, Frye, and Hattie. (2016). Visible Learning. Corwin