Some Data on Baselining

This is a response to Chad’s recent post about baselining, because I’ve found it a useful practice.

I pre-test in all of my intro to American courses (and have posted about it before), using a combination of the citizenship test and questions pulled from the AP Government exam.  My data shows systematic improvement, with all students either matching their previous score or improving, some dramatically.  Last fall the average score on the pretest increased from a 2.9 to a 7.4 out of 10, with an average increase of 3.8 points over 18 students.  This spring the increase was a 3.2 to a 7.0, with the average increase at 3.75 over 22 students.  I need to look closer at the data to see if there is testing bias, but I’m fairly pleased with the results.

I find pre- and post- testing useful for a number of reasons.  First, it provides a nice frame for the course.  As I’ve discussed previously, starting off a course with a discussion about citizenship requirements and whether or not everyone should have to pass a test to enjoy the benefits of citizenship, and then having students actually taking the official test can start the course with a bang.  You can then return to the conversation at the end of the course, this time asking whether everyone should be required to take such a course before they can vote.

Second, the course provides a clear indication to the students of whether or not they have learned anything.  Although we review the answers immediately, I do not return the pre-tests until they have finished their post-tests so they can compare their answers and see what they’ve learned over the course of the term.

Third, the data is very useful for assessment.  My course is in our general education program, and I have to provide data annually on whether students are meeting the learning outcomes.  The pre- and post-test model really helps provide concrete data to put in my reports.

Finally, the data is useful as a check on my own teaching.  If there is little improvement, then clearly I am doing something wrong in the classroom.  As I change things around in my classes, adding simulations, papers, projects, and other elements, the testing model lets me get a small sense of whether an element is having an impact on student learning.