Wednesday, January 10, 2018

Teach 180: The Bias Project (Day 82)

"Would you eat this pizza?" Sure.  It looks delicious.  "Would you eat this vegan pizza?"  Maybe not.  But it's the same pizza!?!  Did introducing the word "vegan" in the question bias the results?  This is an example of two questions that were asked as part of an experiment for the bias project.

In my Prob/Stats class, I have students create an experiment where they purposefully see if they can bias the results of a question.  After gathering data, students used a simulation in Fathom to see if the results based on the biased question are likely to happen by chance alone.  Students put their results on posters and I hang the posters around the room.  Then, half the class stands by their posters and they present to the few students that are in front of them.  This takes about five minutes.  Then, the students move on to another poster and pair of presenters.  Students get to give their presentation about three or four times.  No powerpoints, no notecards, less nervousness and similar to informal presentations they might need to make at some point in their lives in the future.

I told my husband about this project this morning and he asked if my students peer review the posters as part of assigning a grade.  When I asked him how I would go about this, he said to create a rubric.  I am ok with creating rubrics, but I often feel that they can be challenging to have students (and adults) use the rubrics in the same way.  By this I mean that it is hard to have evaluators get similar results with a rubric unless there is extensive training. (AP Statistics teachers who have been to the AP Stat reading can attest to this.)  So, instead of a rubric I had students give qualitative feedback.  For each presentation they saw, they had to write a "Kudos" and an "Improvement".  As I read through the responses this afternoon, I felt like the students gave honest and helpful feedback to each other.  Although I am a math and numbers person, qualitative feedback can often be more powerful than a single number.

No comments:

Post a Comment