How to motivate students to take a "gauge" test that helps us with statistics but does not contribute to their grade?

Question

I am TA-ing for a college class of general physics with currently ~60 enrolled students. We initially (at the beginning of each semester) give them the FCI test ("force concept inventory") that gauges their knowledge/intuition of forces and underlying Newton's laws, and then we repeatedly issue the test at the end of the semester to gain some statistics of how much they learned during the class. The goal is to have some standardized test that probes the gain of knowledge during the semester, so judging just by the final or midterm is not possible (as those cover different topics in physics each time). The test is usually not incentivized by any points or grade at the beginning of the class, so the incentive to take it at the end is a question we face every time.

They know this test doesn't contribute to their grades, so not anyone might feel compelled to take it (this is somewhat similar to the course evaluations). However, for the statistics to be accurate, all those still enrolled should take it. Consider:

• if only a part of the class takes the test, the statistics might be skewed
• if only students who feel strong in physics take it, this would skew the test upward
• if only those who are weaker in physics take it, it skewes the results downward

We might need to somehow incentivize taking this test, but all solutions that initially seem like a no-brainer, might create a new problem:

• providing small amount of bonus points toward the final grade for participation might incentivize it to students who are doing bad in the class, while those who are doing great (well within the range to get an A) might decide to pass, thus removing the best outliers and skewing the results downward

• providing small amount of bonus points just for the participation might make people take it, but not take it seriously, i.e. totally invalidating the results of the test

• I thought of maybe issuing the test during a normal class time, but this seems to me like cheating the students into taking it

• appealing by our needs to get this important statistics that will help the future students yada yada, can motivate better students into it, but maybe the most disgruntled, overwhelmed and/or tired students might just decide to skip it, skewing the statistics upward.

What should be the wording of an announcement reminding them of this test to get most of the students to take it while not make anyone feel like it's okay to skip it or not take it seriously?

Answer 1

I have always seen this done not by issuing bonus points, but by issuing required points for taking the test, e.g. taking each test is work 5% of your grade. The poor students will do it for the free points, and the good students will do it to keep their grade up.

Don't be too focused on "statistical validity." You're trying to reflect on your teaching, not publish a paper. You can look at who took the test(s) and how well they did, whether there was a bias in who self-selected. After all if the strongest students went from 0.8 to 0.95 on how good they were, there is not much room to examine your teaching: a ceiling effect. You will find room for improvement in whether you are helping the weaker students effectively. They will respond for bonus points.