One way to not assess a 5th grader’s understanding of science—the CRCT.
Today, I had a fascinating conversation on Twitter with a number of colleagues both near and far. It started with this tweet by me to my colleague @adchempages:
I’ll pause here for a bit of commentary. For those that don’t know, the CRCT is the Criterion Referenced Competency Test taken by Georgia public school students as a measure of their comprehension of various state standards. This is the start of something I find wrong with the education system today. I don’t mean to pick on my colleague, since I would say he is like nearly every other parent out there, but I am troubled that parents simply accept CRCT scores as “evidence” that their students are progressing. I also know that I think it’s tied to our culture’s response to information overload and constant searching for short metrics that convey quality, but I think this phenomenon applied to other parts of education (college rankings, SAT scores, the challenge index ranking of high schools, etc) has had a negative impact on education, could CRCT scores be doing the same? Would education improve if parents insisted to know what is on the exam and exactly how their child performed, beyond the single number summary they are given?
So I then posted this:
I’ve included a copy of the CRCT practice test here so you can get an idea of its content:
And then the twitter fun begins, with more people adding their thoughts:
This is one of my big points—I don’t think this test really is testing science—it’s testing whether or not the student has learned the names of a bunch of phenomena he/she does not understand “static electricity”,”current electricity,” etc.
Richard Feynman describes the value of this type of learning better than almost anyone else (be sure to listen to the whole video to hear a real example of scientific learning):
This point is certainly vaild, and I didn’t really address it in the conversation. I think my problem with this test isn’t so much that it presents ideas that aren’t really correct with a high school understanding of physics (though I dislike this), my problem is more that it seems to push students toward simply getting the name right, without any desire to test whether or not the student knows how that particular fact was discovered, or how to apply that fact in a new context.
In response to this tweet, yes, I would prefer a lab exam. I’d prefer giving a child a battery, bulb and a single wire, and having him/her light the bulb. Since this task routinely takes my students 20 minutes or more (and I doubt I have any kids who were ever in any danger of not passing the 5th grade science CRCT, I think this may indicate that these students never fully understood these ideas.
And this is something I am very curious about. How would my 9th grade students do on this CRCT practice exam?
I DM’d a link to the test to @BHS_Doyle, who is helping to set elementary school standards in New Jersey, and has written a number of wonderful posts on the topic here.
Here was his response.
ow–test is exactly the kind of stuff that kills science. Now, if the child had the battery/wires/etc/ in front when taking the test….
And this is the sort of exam that was involved in the CRCT cheating scandal at APS. While I find what those the teachers did to be involved unethical, I still have to look at this meaningless exam, and try to see things a bit from their perspective. Lots of research shows students cheat most where there is no connection between teacher and student, and students have no clear understaning of the intrinsic value of the assessment. Did APS teachers have this? My guess is no, and so the treshold for resorting to cheering became all the more closer.
So how would you win this back? Are there good standardized ways to measure learning in elementary school?
Frank Noschese informed me that not every elementary school standardized test needs to look like the one above. The state of New York administers a performance based standardized assessment to all of 5th graders. One example is having students determine what variables affect how far a golf ball moves a cup when it is rolled down a ramp.